7. Expressions

Syntax:

expression:
    primary-expression
    bitwise-expression
    logical-expression
    comparison-expression
    additive-expression
    multiplicative-expression

dash: one of
    - (U+002D)
    EnDash character (U+2013)
    EmDash character (U+2014)
    Horizontal bar character (U+2015)

dashdash:
    dash dash

Description:

An expression is a sequence of operators and operands that designates a method, a function, a writable location, or a value; specifies the computation of a value; produces one or more side effects; or performs some combination thereof. For example,

  • The literal 123 is an expression that designates the int value 123.
  • The expression 1,2,3,4 designates the 4-element array object having the values shown.
  • The expression 10.4 * $a specifies a computation.
  • The expression $a++ produces a side effect.
  • The expression $a[$i--] = $b[++$j] performs a combination of these things.

Except as specified for some operators, the order of evaluation of terms in an expression and the order in which side effects take place are both unspecified. Examples of unspecified behavior include the following: $i++ + $i, $i + --$i, and $w[$j++] = $v[$j].

An implementation of PowerShell may provide support for user-defined types, and those types may have operations defined on them. All details of such types and operations are implementation defined.

A top-level expression is one that is not part of some larger expression. If a top-level expression contains a side-effect operator the value of that expression is not written to the pipeline; otherwise, it is. See §7.1.1 for a detailed discussion of this.

Ordinarily, an expression that designates a collection ([§4§4]) is enumerated into its constituent elements when the value of that expression is used. However, this is not the case when the expression is a cmdlet invocation. For example,

$x = 10,20,30
$a = $($x; 99)                     # $a.Length is 4

$x = New-Object 'int[]' 3
$a = $($x; 99)                     # equivalent, $a.Length is 4

$a = $(New-Object 'int[]' 3; 99)   # $a.Length is 2

In the first two uses of the $(...) operator, the expression designating the collection is the variable $x, which is enumerated resulting in three int values, plus the int 99. However, in the third case, the expression is a direct call to a cmdlet, so the result is not enumerated, and $a is an array of two elements, int[3] and int.

If an operation is not defined by PowerShell, the type of the value designated by the left operand is inspected to see if it has a corresponding op_<operation> method.

7.1 Primary expressions

Syntax:

primary-expression:
    value
    member-access
    element-access
    invocation-expression
    post-increment-expression
    post-decrement-expression

value:
    parenthesized-expression
    sub-expression
    array-expression
    script-block-expression
    hash-literal-expression
    literal
    type-literal
    variable

7.1.1 Grouping parentheses

Syntax:

Tip

The ~opt~ notation in the syntax definitions indicates that the lexical entity is optional in the syntax.

parenthesized-expression:
    ( new-lines~opt~ pipeline new-lines~opt~ )

Description:

A parenthesized expression is a primary-expression whose type and value are the same as those of the expression without the parentheses. If the expression designates a variable then the parenthesized expression designates that same variable. For example, $x.m and ($x).m are equivalent.

Grouping parentheses may be used in an expression to document the default precedence and associativity within that expression. They can also be used to override that default precedence and associativity. For example,

4 + 6 * 2    # 16
4 + (6 * 2)  # 16 document default precedence
(4 + 6) * 2  # 20 override default precedence

Ordinarily, grouping parentheses at the top-most level are redundant. However, that is not always the case. Consider the following example:

2,4,6       # Length 3; values 2,4,6
(2,4),6     # Length 2; values [object[]],int

In the second case, the parentheses change the semantics, resulting in an array whose two elements are an array of 2 ints and the scalar int 6.

Here's another exception:

23.5/2.4          # pipeline gets 9.79166666666667
$a = 1234 * 3.5   # value not written to pipeline
$a                # pipeline gets 4319

In the first and third cases, the value of the result is written to the pipeline. However, although the expression in the second case is evaluated, the result is not written to the pipeline due to the presence of the side-effect operator = at the top level. (Removal of the $a = part allows the value to be written, as * is not a side-effect operator.)

To stop a value of any expression not containing top-level side effects from being written to the pipeline, discard it explicitly, as follows:

# None of these value are written to pipeline
[void](23.5/2.4)
[void]$a
$null = $a
$a > $null

To write to the pipeline the value of any expression containing top-level side effects, enclose that expression in parentheses, as follows:

($a = 1234 * 3.5) # pipeline gets 4319

As such, the grouping parentheses in this case are not redundant.

In the following example, we have variable substitution (§2.3.5.2) taking place in a string literal:

">$($a = -23)<"    # value not written to pipeline, get ><
">$(($a = -23))<"  # pipeline gets >-23<

In the first case, the parentheses represent a sub-expression's delimiters not grouping parentheses, and as the top-level expression contains a side-effect operator, the expression's value is not written to the pipeline. Of course, the > and < characters are still written.) If grouping parenthesis are added -- as shown in the second case -- writing is enabled.

The following examples each contain top-level side-effect operators:

$a = $b = 0      # value not written to pipeline
$a = ($b = 0)    # value not written to pipeline
($a = ($b = 0))  # pipeline gets 0

++$a             # value not written to pipeline
(++$b)           # pipeline gets 1

$a--             # value not written to pipeline
($b--)           # pipeline gets 1

The use of grouping parentheses around an expression containing no top-level side effects makes those parentheses redundant. For example;

$a      # pipeline gets 0
($a)    # no side effect, so () redundant

Consider the following example that has two side effects, neither of which is at the top level:

12.6 + ($a = 10 - ++$b) # pipeline gets 21.6.

The result is written to the pipeline, as the top-level expression has no side effects.

7.1.2 Member access

Syntax:

member-access:
    primary-expression . new-line~opt~ member-name
    primary-expression :: new-line~opt~ member-name

Note that no whitespace is allowed after primary-expression.

Description:

The operator . is used to select an instance member from an object, or a key from a Hashtable. The left operand must designate an object, and the right operand must designate an accessible instance member.

Either the right operand designates an accessible instance member within the type of the object designated by the left operand or, if the left operand designates an array, the right operand designates accessible instance members within each element of the array.

Whitespace is not permitted before the . operator.

This operator is left associative.

The operator :: is used to select a static member from a given type. The left operand must designate a type, and the right-hand operand must designate an accessible static member within that type.

Whitespace is not permitted before the :: operator.

This operator is left associative.

If the right-hand operand designates a writable location within the type of the object designated by the left operand, then the whole expression designates a writable location.

Examples:

$a = 10, 20, 30
$a.Length                    # get instance property

(10, 20, 30).Length

$property = "Length"
$a.$property                 # property name is a variable

$h1 = @{ FirstName = "James"; LastName = "Anderson"; IDNum = 123
}
$h1.FirstName                # designates the key FirstName
$h1.Keys                     # gets the collection of keys

[int]::MinValue              # get static property
[double]::PositiveInfinity   # get static property
$property = "MinValue"
[long]::$property            # property name is a variable

foreach ($t in [byte], [int], [long]) {
    $t::MaxValue             # get static property
}

$a = @{ID = 1 }, @{ID = 2 }, @{ID = 3 }
$a.ID                        # get ID from each element in the array

7.1.3 Invocation expressions

Syntax:

invocation-expression:
    primary-expression . new-line~opt~ member-name argument-list
    primary-expression :: new-line~opt~ member-name argument-list

argument-list:
    ( argument-expression-list~opt~ new-lines~opt~ )

Note that no whitespace is allowed after primary-expression.

Description:

An invocation-expression calls the method designated by primary-expression.member-name or primary-expression::member-name. The parentheses in argument-list contain a possibly empty, comma-separated list of expressions that designate the arguments whose values are passed to the method. Before the method is called, the arguments are evaluated and converted according to the rules of §6, if necessary, to match the types expected by the method. The order of evaluation of primary-expression.member-name, primary-expression::member-name, and the arguments is unspecified.

This operator is left associative.

The type of the result of an invocation-expression is a method-designator (§4.5.24).

Examples:

[math]::Sqrt(2.0)            # call method with argument 2.0
[char]::IsUpper("a")         # call method
$b = "abc#$%XYZabc"
$b.ToUpper()                 # call instance method

[math]::Sqrt(2)              # convert 2 to 2.0 and call method
[math]::Sqrt(2D)             # convert 2D to 2.0 and call method
[math]::Sqrt($true)          # convert $true to 1.0 and call method
[math]::Sqrt("20")           # convert "20" to 20 and call method

$a = [math]::Sqrt            # get method descriptor for Sqrt
$a.Invoke(2.0)               # call Sqrt via the descriptor
$a = [math]::("Sq"+"rt")     # get method descriptor for Sqrt
$a.Invoke(2.0)               # call Sqrt via the descriptor
$a = [char]::ToLower         # get method descriptor for ToLower
$a.Invoke("X")               # call ToLower via the descriptor

7.1.4 Element access

Syntax:

element-access:
    primary-expression [ new-lines~opt~ expression new-lines~opt~ ]

Description:

There must not be any whitespace between primary-expression and the left square bracket ([).

7.1.4.1 Subscripting an array

Description:

Arrays are discussed in detail in §9. If expression is a 1-dimensional array, see §7.1.4.5.

When primary-expression designates a 1-dimensional array A, the operator [] returns the element located at A[0 + expression] after the value of expression has been converted to int. The result has the element type of the array being subscripted. If expression is negative, A[expression] designates the element located at A[A.Length + expression].

When primary-expression designates a 2-dimensional array B, the operator [] returns the element located at B[0 + row,0 + column] after the value of the row and column components of expression (which are specified as a comma-separated list) have been converted to int. The result has the element type of the array being subscripted. Unlike for a 1-dimensional array, negative positions have no special meaning.

When primary-expression designates an array of three or more dimensions, the rules for 2-dimensional arrays apply and the dimension positions are specified as a comma-separated list of values.

If a read access on a non-existing element is attempted, the result is $null. It is an error to write to a non-existing element.

For a multidimensional-array subscript expression, the order of evaluation of the dimension position expressions is unspecified. For example, given a 3-dimensional array $a, the behavior of $a[$i++,$i,++$i] is unspecified.

If expression is an array, see §7.1.4.5.

This operator is left associative.

Examples:

$a = [int[]](10,20,30) # [int[]], Length 3
$a[1] # returns int 20
$a[20] # no such position, returns $null
$a[-1] # returns int 30, i.e., $a[$a.Length-1]
$a[2] = 5 # changes int 30 to int 5
$a[20] = 5 # implementation-defined behavior

$a = New-Object 'double[,]' 3,2
$a[0,0] = 10.5 # changes 0.0 to 10.5
$a[0,0]++ # changes 10.5 to 10.6

$list = ("red",$true,10),20,(1.2, "yes")
$list[2][1] # returns string "yes"

$a = @{ A = 10 },@{ B = $true },@{ C = 123.45 }
$a[1]["B"] # $a[1] is a Hashtable, where B is a key

$a = "red","green"
$a[1][4] # returns string "n" from string in $a[1]

If a write access to a non-existing element is attempted, an IndexOutOfRange exception is raised.

7.1.4.2 Subscripting a string

Description:

When primary-expression designates a string S, the operator [] returns the character located in the zero-based position indicated by expression, as a char. If expression is greater than or equal to that string's length, the result is $null. If expression is negative, S[expression] designates the element located at S[S.Length + expression].

Examples:

$s = "Hello"   # string, Length 5, positions 0-4
$c = $s[1]     # returns "e" as a string
$c = $s[20]    # no such position, returns $null
$c = $s[-1]    # returns "o", i.e., $s[$s.Length-1]

7.1.4.3 Subscripting a Hashtable

Description:

When primary-expression designates a Hashtable, the operator [] returns the value(s) associated with the key(s) designated by expression. The type of expression is not restricted.

When expression is a single key name, the result is the associated value and has that type, unless no such key exists, in which case, the result is $null. If $null is used as the key the behavior is implementation defined. If expression is an array of key names, see §7.1.4.5.

If expression is an array, see §7.1.4.5.

Examples:

$h1 = @{ FirstName = "James"; LastName = "Anderson"; IDNum = 123 }
$h1['FirstName']     # the value associated with key FirstName
$h1['BirthDate']     # no such key, returns $null

$h1 = @{ 10 = "James"; 20.5 = "Anderson"; $true = 123 }
$h1[10]              # returns value "James" using key 10
$h1[20.5]            # returns value "Anderson" using key 20.5
$h1[$true]           # returns value 123 using key $true

When expression is a single key name, if $null is used as the only value to subscript a Hashtable, a NullArrayIndex exception is raised.

7.1.4.4 Subscripting an XML document

Description:

When primary-expression designates an object of type xml, expression is converted to string, if necessary, and the operator [] returns the first child element having the name specified by expression. The type of expression must be string. The type of the result is implementation defined. The result can be subscripted to return its first child element. If no child element exists with the name specified by expression, the result is $null. The result does not designate a writable location.

Examples:

$x = [xml]@"
<Name>
<FirstName>Mary</FirstName>
<LastName>King</LastName>
</Name>
"@

$x['Name']                # refers to the element Name
$x['Name']['FirstName']   # refers to the element FirstName within Name
$x['FirstName']           # No such child element at the top level, result is $null

The type of the result is System.Xml.XmlElement or System.String.

7.1.4.5 Generating array slices

When primary-expression designates an object of a type that is enumerable (§4) or a Hashtable, and expression is a 1-dimensional array, the result is an array slice (§9.9) containing the elements of primary-expression designated by the elements of expression.

In the case of a Hashtable, the array slice contains the associated values to the keys provided, unless no such key exists, in which case, the corresponding element is $null. If $null is used as any key name the behavior is implementation defined.

Examples:

$a = [int[]](30,40,50,60,70,80,90)
$a[1,3,5]                 # slice has Length 3, value 40,60,80
$a[,5]                    # slice with Length 1
$a[@()]                   # slice with Length 0
$a[-1..-3]                # slice with Length 3, value 90,80,70
$a = New-Object 'int[,]' 3,2
$a[0,0] = 10; $a[0,1] = 20; $a[1,0] = 30
$a[1,1] = 40; $a[2,0] = 50; $a[2,1] = 60
$a[(0,1),(1,0)]           # slice with Length 2, value 20,30, parens needed
$h1 = @{ FirstName = "James"; LastName = "Anderson"; IDNum = 123 }
$h1['FirstName']          # the value associated with key FirstName
$h1['BirthDate']          # no such key, returns $null
$h1['FirstName','IDNum']  # returns [object[]], Length 2 (James/123)
$h1['FirstName','xxx']    # returns [object[]], Length 2 (James/$null)
$h1[$null,'IDNum']        # returns [object[]], Length 2 ($null/123)

Windows PowerShell: When expression is a collection of two or more key names, if $null is used as any key name that key is ignored and has no corresponding element in the resulting array.

7.1.5 Postfix increment and decrement operators

Syntax:

post-increment-expression:
    primary-expression ++

post-decrement-expression:
    primary-expression dashdash

Description:

The primary-expression must designate a writable location having a value of numeric type (§4) or the value $null. If the value designated by the operand is $null, that value is converted to type int and value zero before the operator is evaluated. The type of the value designated by primary-expression may change when the result is stored. See §7.11 for a discussion of type change via assignment.

The result produced by the postfix ++ operator is the value designated by the operand. After that result is obtained, the value designated by the operand is incremented by 1 of the appropriate type. The type of the result of expression E++ is the same as for the result of the expression E + 1 (§7.7).

The result produced by the postfix -- operator is the value designated by the operand. After that result is obtained, the value designated by the operand is decremented by 1 of the appropriate type. The type of the result of expression E-- is the same as for the result of the expression E - 1 (§7.7).

These operators are left associative.

Examples:

$i = 0                # $i = 0
$i++                  # $i is incremented by 1
$j = $i--             # $j takes on the value of $i before the decrement

$a = 1,2,3
$b = 9,8,7
$i = 0
$j = 1
$b[$j--] = $a[$i++]   # $b[1] takes on the value of $a[0], then $j is
                      # decremented, $i incremented

$i = 2147483647       # $i holds a value of type int
$i++                  # $i now holds a value of type double because
                      # 2147483648 is too big to fit in type int

[int]$k = 0           # $k is constrained to int
$k = [int]::MaxValue  # $k is set to 2147483647
$k++                  # 2147483648 is too big to fit, imp-def behavior

$x = $null            # target is unconstrained, $null goes to [int]0
$x++                  # value treated as int, 0->1

7.1.6 $(...) operator

Syntax:

sub-expression:
    $( new-lines~opt~ statement-list~opt~ new-lines~opt~ )

Description:

If statement-list is omitted, the result is $null. Otherwise, statement-list is evaluated. Any objects written to the pipeline as part of the evaluation are collected in an unconstrained 1-dimensional array, in order. If the array of collected objects is empty, the result is $null. If the array of collected objects contains a single element, the result is that element; otherwise, the result is the unconstrained 1-dimensional array of collected results.

Examples:

$j = 20
$($i = 10) # pipeline gets nothing
$(($i = 10)) # pipeline gets int 10
$($i = 10; $j) # pipeline gets int 20
$(($i = 10); $j) # pipeline gets [object[]](10,20)
$(($i = 10); ++$j) # pipeline gets int 10
$(($i = 10); (++$j)) # pipeline gets [object[]](10,22)
$($i = 10; ++$j) # pipeline gets nothing
$(2,4,6) # pipeline gets [object[]](2,4,6)

7.1.7 @(...) operator

Syntax:

array-expression:
    @( new-lines~opt~ statement-list~opt~ new-lines~opt~ )

Description:

If statement-list is omitted, the result is an unconstrained 1-dimensional array of length zero. Otherwise, statement-list is evaluated, and any objects written to the pipeline as part of the evaluation are collected in an unconstrained 1-dimensional array, in order. The result is the (possibly empty) unconstrained 1-dimensional array.

Examples:

$j = 20
@($i = 10)             # 10 not written to pipeline, result is array of 0
@(($i = 10))           # pipeline gets 10, result is array of 1
@($i = 10; $j)         # 10 not written to pipeline, result is array of 1
@(($i = 10); $j)       # pipeline gets 10, result is array of 2
@(($i = 10); ++$j)     # pipeline gets 10, result is array of 1
@(($i = 10); (++$j))   # pipeline gets both values, result is array of 2
@($i = 10; ++$j)       # pipeline gets nothing, result is array of 0

$a = @(2,4,6)          # result is array of 3
@($a)                  # result is the same array of 3
@(@($a))               # result is the same array of 3

7.1.8 Script block expression

Syntax:

script-block-expression:
    { new-lines~opt~ script-block new-lines~opt~ }

script-block:
    param-block~opt~ statement-terminators~opt~ script-block-body~opt~

script-block-body:
    named-block-list
    statement-list

Description:

param-block is described in §8.10.9. named-block-list is described in §8.10.7.

A script block is an unnamed block of statements that can be used as a single unit. Script blocks can be used to invoke a block of code as if it was a single command, or they can be assigned to variables that can be executed.

The named-block-list or statement-list is executed and the type and value(s) of the result are the type and value(s) of the results of those statement sets.

A script-block-expression has type scriptblock (§4.3.7).

If param-block is omitted, any arguments passed to the script block are available via $args (§8.10.1).

During parameter binding, a script block can be passed either as a script block object or as the result after the script block has been evaluated. See §6.17 for further information.

7.1.9 Hash literal expression

Syntax:

hash-literal-expression:
    @{ new-lines~opt~ hash-literal-body~opt~ new-lines~opt~ }

hash-literal-body:
    hash-entry
    hash-literal-body statement-terminators hash-entry

hash-entry:
    key-expression = new-lines~opt~ statement

key-expression:
    simple-name
    unary-expression

statement-terminators:
    statement-terminator
    statement-terminators statement-terminator

statement-terminator:
    ;
    new-line-character

Description:

A hash-literal-expression is used to create a Hashtable (§10) of zero or more elements each of which is a key/value pair.

The key may have any type except the null type. The associated values may have any type, including the null type, and each of those values may be any expression that designates the desired value, including $null.

The ordering of the key/value pairs is not significant.

Examples:

$h1 = @{ FirstName = "James"; LastName = "Anderson"; IDNum = 123 }
$last = "Anderson"; $IDNum = 120
$h2 = @{ FirstName = "James"; LastName = $last; IDNum = $IDNum + 3 }
$h3 = @{ }
$h4 = @{ 10 = "James"; 20.5 = "Anderson"; $true = 123 }

which creates two Hashtables, $h1 and $h2, each containing three key/value pairs, and a third, $h3, that is empty. Hashtable $h4 has keys of various types.

7.1.10 Type literal expression

Syntax:

type-literal:
    [ type-spec ]

type-spec:
    array-type-name new-lines~opt~ dimension~opt~ ]
    generic-type-name new-lines~opt~ generic-type-arguments ]
    type-name

dimension:
    ,
    dimension ,

generic-type-arguments:
    type-spec new-lines~opt~
    generic-type-arguments , new-lines~opt~ type-spec

array-type-name:
    type-name [

generic-type-name:
    type-name [

Description:

A type-literal is represented in an implementation by some unspecified underlying type. As a result, a type name is a synonym for its underlying type.

Type literals are used in a number of contexts:

  • Specifying an explicit conversion (§6, §7.2.9)
  • Creating a type-constrained array (§9.4)
  • Accessing the static members of an object (§7.1.2)
  • Specifying a type constraint on a variable (§5.3) or a function parameter (§8.10.2)

Examples:

[int].IsPrimitive        # $true
[Object[]].FullName      # "System.Object[]"
[int[,,]].GetArrayRank() # 3

A generic stack type (§4.4) that is specialized to hold strings might be written as [Stack[string]], and a generic dictionary type that is specialized to hold int keys with associated string values might be written as [Dictionary[int,string]].

The type of a type-literal is System.Type. The complete name for the type Stack[string] suggested above is System.Collections.Generic.Stack[int]. The complete name for the type Dictionary[int,string] suggested above is System.Collections.Generic.Dictionary[int,string].

7.2 Unary operators

Syntax:

unary-expression:
    primary-expression
    expression-with-unary-operator

expression-with-unary-operator:
    , new-lines~opt~ unary-expression
    -not new-lines~opt~ unary-expression
    ! new-lines~opt~ unary-expression
    -bnot new-lines~opt~ unary-expression
    + new-lines~opt~ unary-expression
    dash new-lines~opt~ unary-expression
    pre-increment-expression
    pre-decrement-expression
    cast-expression
    -split new-lines~opt~ unary-expression
    -join new-lines~opt~ unary-expression

pre-increment-expression:
    ++ new-lines~opt~ unary-expression

pre-decrement-expression:
    dashdash new-lines~opt~ unary-expression

dash: one of
    - (U+002D)
    EnDash character (U+2013)
    EmDash character (U+2014)
    Horizontal bar character (U+2015)

cast-expression:
    type-literal unary-expression

dashdash:
    dash dash

7.2.1 Unary comma operator

Description:

The comma operator (,) creates an unconstrained 1-dimensional array having one element, whose type and value are that of unary-expression.

This operator is right associative.

Examples:

$a = ,10         # create an unconstrained array of 1 element, $a[0],
                 # which has type int

$a = ,(10,"red") # create an unconstrained array of 1 element,
$a[0],
                 # which is an unconstrained array of 2 elements,
                 # $a[0][0] an int, and $a[0][1] a string

$a = ,,10        # create an unconstrained array of 1 element, which is
                 # an unconstrained array of 1 element, which is an int
                 # $a[0][0] is the int. Contrast this with @(@(10))

7.2.2 Logical NOT

Syntax:

logical-not-operator:
    dash not

dash: one of
    - (U+002D)
    EnDash character (U+2013)
    EmDash character (U+2014)
    Horizontal bar character (U+2015)

Description:

The operator -not converts the value designated by unary-expression to type bool (§6.2), if necessary, and produces a result of that type. If unary-expression's value is True, the result is False, and vice versa. The operator ! is an alternate spelling for -not.

This operator is right associative.

Examples:

-not $true         # False
-not -not $false   # False
-not 0             # True
-not 1.23          # False
!"xyz"             # False

7.2.3 Bitwise NOT

Syntax:

bitwise-not-operator:
    dash bnot

dash: one of
    - (U+002D)
    EnDash character (U+2013)
    EmDash character (U+2014)
    Horizontal bar character (U+2015)

Description:

The operator -bnot converts the value designated by unary-expression to an integer type (§6.4), if necessary. If the converted value can be represented in type int then that is the result type. Else, if the converted value can be represented in type long then that is the result type. Otherwise, the expression is ill-formed. The resulting value is the ones-complement of the converted value.

This operator is right associative.

Examples:

-bnot $true         # int with value 0xFFFFFFFE
-bnot 10            # int with value 0xFFFFFFF5
-bnot 2147483648.1  # long with value 0xFFFFFFFF7FFFFFFF
-bnot $null         # int with value 0xFFFFFFFF
-bnot "0xabc"       # int with value 0xFFFFF543

7.2.4 Unary plus

Description:

An expression of the form + unary-expression is treated as if it were written as 0 + unary-expression (§7.7). The integer literal 0 has type int.

This operator is right associative.

Examples:

+123L         # type long, value 123
+0.12340D     # type decimal, value 0.12340
+"0xabc"      # type int, value 2748

7.2.5 Unary minus

Description:

An expression of the form - unary-expression is treated as if it were written as 0 - unary-expression (§7.7). The integer literal 0 has type int. The minus operator can be any one of the dash characters listed in §7.2.

This operator is right associative.

Examples:

-$true     # type int, value -1
-123L      # type long, value -123
-0.12340D  # type decimal, value -0.12340

7.2.6 Prefix increment and decrement operators

Description:

The unary-expression must designate a writable location having a value of numeric type (§4) or the value $null. If the value designated by its unary-expression is $null, unary-expression's value is converted to type int and value zero before the operator is evaluated.

Note

The type of the value designated by unary-expression may change when the result is stored. See §7.11 for a discussion of type change via assignment.

For the prefix increment operator ++ , the value of unary-expression is incremented by 1 of the appropriate type. The result is the new value after incrementing has taken place. The expression ++E is equivalent to E += 1 (§7.11.2).

For the prefix decrement operator -- , the value of unary-expression is decremented by 1 of the appropriate type. The result is the new value after decrementing has taken place. The expression --E is equivalent to E -= 1 (§7.11.2). The prefix decrement operator can be any of the patterns matching the dashdash pattern in §7.2.

These operators are right associative.

Examples:

$i = 0                # $i = 0
++$i                  # $i is incremented by 1
$j = --$i             # $i is decremented then $j takes on the value of $i

$a = 1,2,3
$b = 9,8,7
$i = 0;
$j = 1
$b[--$j] = $a[++$i]   # $j is # decremented, $i incremented, then $b[0]
                      # takes on the value of $a[1]

$i = 2147483647       # $i holds a value of type int
++$i                  # $i now holds a value of type double because
                      # 2147483648 is too big to fit in type int

[int]$k = 0           # $k is constrained to int
$k = [int]::MinValue  # $k is set to -2147483648
--$k                  # -2147483649 is too small to fit, imp-def behavior

$x = $null            # target is unconstrained, $null goes to [int]0
--$x                  # value treated as int, 0 becomes -1

7.2.7 The unary -join operator

Syntax:

join-operator:
    dash join

dash: one of
    - (U+002D)
    EnDash character (U+2013)
    EmDash character (U+2014)
    Horizontal bar character (U+2015)

Description:

The unary -join operator produces a string that is the concatenation of the value of one or more objects designated by unary-expression. (A separator can be inserted by using the binary version of this operator (§7.8.4.4).)

unary-expression can be a scalar value or a collection.

Examples:

-join (10, 20, 30)             # result is "102030"
-join (123, $false, 19.34e17)  # result is "123False1.934E+18"
-join 12345                    # result is "12345"
-join $null                    # result is ""

7.2.8 The unary -split operator

Syntax:

split-operator:
    dash split

dash: one of
    - (U+002D)
    EnDash character (U+2013)
    EmDash character (U+2014)
    Horizontal bar character (U+2015)

Description:

The unary -split operator splits one or more strings designated by unary-expression, returning their subparts in a constrained 1-dimensional array of string. It treats any contiguous group of whitespace characters as the delimiter between successive subparts. An explicit delimiter string can be specified by using the binary version of this operator (§7.8.4.5) or its two variants (§7.8).

The delimiter text is not included in the resulting strings. Leading and trailing whitespace in the input string is ignored. An input string that is empty or contains whitespace only results in an array of one string, which is empty.

unary-expression can designate a scalar value or an array of strings.

Examples:

-split " red`tblue`ngreen " # 3 strings: "red", "blue", "green"
-split ("yes no", "up down") # 4 strings: "yes", "no", "up", "down"
-split " " # 1 (empty) string

7.2.9 Cast operator

Description:

This operator converts explicitly (§6) the value designated by unary-expression to the type designated by type-literal (§7.1.10). If type-literal is other than void, the type of the result is the named type, and the value is the value after conversion. If type-literal is void, no object is written to the pipeline and there is no result.

When an expression of any type is cast to that same type, the resulting type and value is the unary-expression's type and value.

This operator is right associative.

Examples:

[bool]-10        # a bool with value True
[int]-10.70D     # a decimal with value -10
[int]10.7        # an int with value 11
[long]"+2.3e+3"  # a long with value 2300
[char[]]"Hello"  # an array of 5 char with values H, e, l, l, and o.

7.3 Binary comma operator

Syntax:

array-literal-expression:
    unary-expression , new-lines~opt~ array-literal-expression

Description:

The binary comma operator creates a 1-dimensional array whose elements are the values designated by its operands, in lexical order. The array has unconstrained type.

Examples:

2,4,6                    # Length 3; values 2,4,6
(2,4),6                  # Length 2; values [object[]],int
(2,4,6),12,(2..4)        # Length 3; [object[]],int,[object[]]
2,4,6,"red",$null,$true  # Length 6

The addition of grouping parentheses to certain binary comma expressions does not document the default precedence; instead, it changes the result.

7.4 Range operator

Syntax:

range-expression:
    unary-expression .. new-lines~opt~ unary-expression

Description:

A range-expression creates an unconstrained 1-dimensional array whose elements are the values of the int sequence specified by the range bounds. The values designated by the operands are converted to int, if necessary (§6.4). The operand designating the lower value after conversion is the lower bound, while the operand designating the higher value after conversion is the upper bound. Both bounds may be the same, in which case, the resulting array has length 1. If the left operand designates the lower bound, the sequence is in ascending order. If the left operand designates the upper bound, the sequence is in descending order.

Conceptually, this operator is a shortcut for the corresponding binary comma operator sequence. For example, the range 5..8 can also be generated using 5,6,7,8. However, if an ascending or descending sequence is needed without having an array, an implementation may avoid generating an actual array. For example, in foreach ($i in 1..5) { ... }, no array need be created.

A range-expression can be used to specify an array slice (§9.9).

Examples:

1..10        # ascending range 1..10
-495..-500   # descending range -495..-500
16..16       # sequence of 1

$x = 1.5
$x..5.40D    # ascending range 2..5

$true..3     # ascending range 1..3
-2..$null    # ascending range -2..0
0xf..0xa     # descending range 15..10

7.5 Format operator

Syntax:

format-expression:
    format-specification-string format-operator new-lines~opt~ range-expression

format-operator:
    dash f

dash:
    - (U+002D)
    EnDash character (U+2013)
    EmDash character (U+2014)
    Horizontal bar character (U+2015)

Description:

A format-expression formats one or more values designated by range-expression according to a format-specification-string designated by format-expression. The positions of the values designated by range-expression are numbered starting at zero and increasing in lexical order. The result has type string.

A format specification string may contain zero or more format specifications each having the following form:

{N [ ,M ][ : FormatString ]}

N represents a (required) range-expression value position, M represents the (optional) minimum display width, and FormatString indicates the (optional) format. If the width of a formatted value exceeds the specified width, the width is increased accordingly. Values whose positions are not referenced in FormatString are ignored after being evaluated for any side effects. If N refers to a non-existent position, the behavior is implementation defined. Value of type $null and void are formatted as empty strings. Arrays are formatted as for sub-expression (§7.1.6). To include the characters { and } in a format specification without their being interpreted as format delimiters, write them as {{ and }}, respectively.

For a complete definition of format specifications, see the type System.IFormattable in Ecma Technical Report TR/84.

Examples:

"__{0,3}__" -f 5                         # __ 5__
"__{0,-3}__" -f 5                        # __5 __
"__{0,3:000}__" -f 5                     # __005__
"__{0,5:0.00}__" -f 5.0                  # __ 5.00__
"__{0:C}__" -f 1234567.888               # __$1,234,567.89__
"__{0:C}__" -f -1234.56                  # __($1,234.56)__
"__{0,12:e2}__" -f 123.456e2             # __ 1.23e+004__
"__{0,-12:p}__" -f -0.252                # __-25.20 % __

$i = 5; $j = 3
"__{0} + {1} <= {2}__" -f $i,$j,($i+$j)  # __5 + 3 <= 8__

$format = "__0x{0:X8}__"
$format -f 65535                         # __0x0000FFFF__

In a format specification, if N refers to a non-existent position, a FormatError is raised.

7.6 Multiplicative operators

Syntax:

multiplicative-expression:
    multiplicative-expression * new-lines~opt~ format-expression
    multiplicative-expression / new-lines~opt~ format-expression
    multiplicative-expression % new-lines~opt~ format-expression

7.6.1 Multiplication

Description:

The result of the multiplication operator * is the product of the values designated by the two operands after the usual arithmetic conversions (§6.15) have been applied.

This operator is left associative.

Examples:

12 * -10L      # long result -120
-10.300D * 12  # decimal result -123.600
10.6 * 12      # double result 127.2
12 * "0xabc"   # int result 32976

7.6.2 String replication

Description:

When the left operand designates a string the binary * operator creates a new string that contains the one designated by the left operand replicated the number of times designated by the value of the right operand as converted to integer type (§6.4).

This operator is left associative.

Examples:

"red" * "3"       # string replicated 3 times
"red" * 4         # string replicated 4 times
"red" * 0         # results in an empty string
"red" * 2.3450D   # string replicated twice
"red" * 2.7       # string replicated 3 times

7.6.3 Array replication

Description:

When the left operand designates an array the binary * operator creates a new unconstrained 1-dimensional array that contains the value designated by the left operand replicated the number of times designated by the value of the right operand as converted to integer type (§6.4). A replication count of zero results in an array of length 1. If the left operand designates a multidimensional array, it is flattened (§9.12) before being used.

This operator is left associative.

Examples:

$a = [int[]](10,20)              # [int[]], Length 2*1
$a * "3"                         # [object[]], Length 2*3
$a * 4                           # [object[]], Length 2*4
$a * 0                           # [object[]], Length 2*0
$a * 2.3450D                     # [object[]], Length 2*2
$a * 2.7                         # [object[]], Length 2*3
(New-Object 'float[,]' 2,3) * 2  # [object[]], Length 2*2

7.6.4 Division

Description:

The result of the division operator / is the quotient when the value designated by the left operand is divided by the value designated by the right operand after the usual arithmetic conversions (§6.15) have been applied.

If an attempt is made to perform integer or decimal division by zero, an implementation-defined terminating error is raised.

This operator is left associative.

Examples:

10/-10      # int result -1
12/-10      # double result -1.2
12/-10D     # decimal result 1.2
12/10.6     # double result 1.13207547169811
12/"0xabc"  # double result 0.00436681222707424

If an attempt is made to perform integer or decimal division by zero, a RuntimeException exception is raised.

7.6.5 Remainder

Description:

The result of the remainder operator % is the remainder when the value designated by the left operand is divided by the value designated by the right operand after the usual arithmetic conversions (§6.15) have been applied.

If an attempt is made to perform integer or decimal division by zero, an implementation-defined terminating error is raised.

Examples:

10 % 3          # int result 1
10.0 % 0.3      # double result 0.1
10.00D % "0x4"  # decimal result 2.00

If an attempt is made to perform integer or decimal division by zero, a RuntimeException exception is raised.

7.7 Additive operators

Syntax:

additive-expression:
    primary-expression + new-lines~opt~ expression
    primary-expression dash new-lines~opt~ expression

dash: one of
    - (U+002D)
    EnDash character (U+2013)
    EmDash character (U+2014)
    Horizontal bar character (U+2015)

7.7.1 Addition

Description:

The result of the addition operator + is the sum of the values designated by the two operands after the usual arithmetic conversions (§6.15) have been applied.

This operator is left associative.

Examples:

12 + -10L       # long result 2
-10.300D + 12   # decimal result 1.700
10.6 + 12       # double result 22.6
12 + "0xabc"    # int result 2760

7.7.2 String concatenation

Description:

When the left operand designates a string the binary + operator creates a new string that contains the value designated by the left operand followed immediately by the value(s) designated by the right operand as converted to type string (§6.8).

This operator is left associative.

Examples:

"red" + "blue"      # "redblue"
"red" + "123"       # "red123"
"red" + 123         # "red123"
"red" + 123.456e+5  # "red12345600"
"red" + (20,30,40)  # "red20 30 40"

7.7.3 Array concatenation

Description:

When the left operand designates an array the binary + operator creates a new unconstrained 1-dimensional array that contains the elements designated by the left operand followed immediately by the value(s) designated by the right operand. Multidimensional arrays present in either operand are flattened (§9.12) before being used.

This operator is left associative.

Examples:

$a = [int[]](10,20)               # [int[]], Length 2
$a + "red"                        # [object[]], Length 3
$a + 12.5,$true                   # [object[]], Length 4
$a + (New-Object 'float[,]' 2,3)  # [object[]], Length 8
(New-Object 'float[,]' 2,3) + $a  # [object[]], Length 8

7.7.4 Hashtable concatenation

Description:

When both operands designate Hashtables the binary + operator creates a new Hashtable that contains the elements designated by the left operand followed immediately by the elements designated by the right operand.

If the Hashtables contain the same key, an implementation-defined terminating error is raised.

This operator is left associative.

Examples:

$h1 = @{ FirstName = "James"; LastName = "Anderson" }
$h2 = @{ Dept = "Personnel" }
$h3 = $h1 + $h2      # new Hashtable, Count = 3

If the Hashtables contain the same key, an exception of type BadOperatorArgument is raised.

7.7.5 Subtraction

Description:

The result of the subtraction operator - is the difference when the value designated by the right operand is subtracted from the value designated by the left operand after the usual arithmetic conversions (§6.15) have been applied. The subtraction operator can be any one of the dash characters listed in §7.7.

This operator is left associative.

Examples:

12 - -10L      # long result 22
-10.300D - 12  # decimal result -22.300
10.6 - 12      # double result -1.4
12 - "0xabc"   # int result -2736

7.8 Comparison operators

Syntax:

comparison-expression:
    primary-expression comparison-operator new-lines~opt~ expression

comparison-operator:
    equality-operator
    relational-operator
    containment-operator
    type-operator
    like-operator
    match-operator

Description:

The type of the value designated by the left operand determines how the value designated by the right operand is converted (§6), if necessary, before the comparison is done.

Some comparison operators have two variants, one that is case sensitive (-c<operator>), and one that isn't case sensitive (-i<operator>). The -<operator> version is equivalent to -i<operator>. Case sensitivity is meaningful only with comparisons of values of type string. In non-string comparison contexts, the two variants behave the same.

These operators are left associative.

7.8.1 Equality and relational operators

Syntax:

equality-operator: one of
    dash eq     dash ceq    dash ieq
    dash ne     dash cne    dash ine

dash:
    - (U+002D)
    EnDash character (U+2013)
    EmDash character (U+2014)
    Horizontal bar character (U+2015)

relational-operator: one of
    dash lt     dash clt    dash ilt
    dash le     dash cle    dash ile
    dash gt     dash cgt    dash igt
    dash ge     dash cge    dash ige

Description:

There are two equality operators: equality (-eq) and inequality (-ne); and four relational operators: less-than (-lt), less-than-or-equal-to (-le), greater-than (-gt), and greater-than-or-equal-to (-ge). Each of these has two variants (§7.8).

For two strings to compare equal, they must have the same length and contents, and letter case, if appropriate.

If the value designated by the left operand is not a collection, the result has type bool. Otherwise, the result is a possibly empty unconstrained 1-dimensional array containing the elements of the collection that test True when compared to the value designated by the right operand.

Examples:

10 -eq "010"           # True, int comparison
"010" -eq 10           # False, string comparison
"RED" -eq "Red"        # True, case-insensitive comparison
"RED" -ceq "Red"       # False, case-sensitive comparison
"ab" -lt "abc"         # True

10,20,30,20,10 -ne 20  # 10,30,10, Length 3
10,20,30,20,10 -eq 40  # Length 0
10,20,30,20,10 -ne 40  # 10,20,30,20,10, Length 5
10,20,30,20,10 -gt 25  # 30, Length 1
0,1,30 -ne $true       # 0,30, Length 2
0,"00" -eq "0"         # 0 (int), Length 1

7.8.2 Containment operators

Syntax:

containment-operator: one of
    dash contains       dash ccontains      dash icontains
    dash notcontains    dash cnotcontains   dash inotcontains
    dash in             dash cin            dash iin
    dash notin          dash cnotin         dash inotin

dash:
    - (U+002D)
    EnDash character (U+2013)
    EmDash character (U+2014)
    Horizontal bar character (U+2015)

Description:

There are four containment operators: contains (-contains), does-not-contain (-notcontains), in (-in) and not-in (-notin). Each of these has two variants (§7.8).

The containment operators return a result of type bool that indicates whether a value occurs (or does not occur) at least once in the elements of an array. With -contains and -notcontains, the value is designated by the right operand and the array is designated by the left operand. With -in and -notin, the operands are reversed. The value is designated by the left operand and the array is designated by the right operand.

For the purposes of these operators, if the array operand has a scalar value, the scalar value is treated as an array of one element.

Examples:

10,20,30,20,10 -contains 20     # True
10,20,30,20,10 -contains 42.9   # False
10,20,30 -contains "10"         # True
"10",20,30 -contains 10         # True
"010",20,30 -contains 10        # False
10,20,30,20,10 -notcontains 15  # True
"Red",20,30 -ccontains "RED"    # False

7.8.3 Type testing and conversion operators

Syntax:

type-operator: one of
    dash is
    dash as

dash:
    - (U+002D)
    EnDash character (U+2013)
    EmDash character (U+2014)
    Horizontal bar character (U+2015)

Description:

The type operator -is tests whether the value designated by the left operand has the type, or is derived from a type that has the type, designated by the right operand. The right operand must designate a type or a value that can be converted to a type (such as a string that names a type). The type of the result is bool. The type operator -isnot returns the logical negation of the corresponding -is form.

The type operator -as attempts to convert the value designated by the left operand to the type designated by the right operand. The right operand must designate a type or a value that can be converted to a type (such as a string that names a type). If the conversion fails, $null is returned; otherwise, the converted value is returned and the return type of that result is the runtime type of the converted value.

Examples:

$a = 10            # value 10 has type int
$a -is [int]       # True

$t = [int]
$a -isnot $t       # False
$a -is "int"       # True
$a -isnot [double] # True

$x = [int[]](10,20)
$x -is [int[]]     # True

$a = "abcd"        # string is derived from object
$a -is [object]    # True

$x = [double]
foreach ($t in [int],$x,[decimal],"string") {
    $b = (10.60D -as $t) * 2  # results in int 22, double 21.2
}                             # decimal 21.20, and string "10.6010.60"

7.8.4 Pattern matching and text manipulation operators

7.8.4.1 The -like and -notlike operators

Syntax:

like-operator: one of
    dash like       dash clike      dash ilike
    dash notlike    dash cnotlike   dash inotlike

dash:
    - (U+002D)
    EnDash character (U+2013)
    EmDash character (U+2014)
    Horizontal bar character (U+2015)

Description:

If the left operand does not designate a collection, the result has type bool. Otherwise, the result is a possibly empty unconstrained 1-dimensional array containing the elements of the collection that test True when compared to the value designated by the right operand. The right operand may designate a string that contains wildcard expressions (§3.15). These operators have two variants (§7.8).

Examples:

"Hello" -like "h*"                   # True, starts with h
"Hello" -clike "h*"                  # False, does not start with lowercase h
"Hello" -like "*l*"                  # True, has an l in it somewhere
"Hello" -like "??l"                  # False, no length match

"-abc" -like "[-xz]*"                # True, - is not a range separator
"#$%\^&" -notlike "*[A-Za-z]"        # True, does not end with alphabetic character
"He" -like "h[aeiou]?*"              # False, need at least 3 characters
"When" -like "*[?]"                  # False, ? is not a wildcard character
"When?" -like "*[?]"                 # True, ? is not a wildcard character

"abc","abbcde","abcgh" -like "abc*"  # object[2], values
"abc" and "abcgh"

7.8.4.2 The -match and -notmatch operators

Syntax:

match-operator: one of
    dash match      dash cmatch     dash imatch
    dash notmatch   dash cnotmatch  dash inotmatch

dash:
    - (U+002D)
    EnDash character (U+2013)
    EmDash character (U+2014)
    Horizontal bar character (U+2015)

Description:

If the left operand does not designate a collection, the result has type bool and if that result is $true, the elements of the Hashtable $matches are set to the strings that match (or do-not-match) the value designated by the right operand. Otherwise, the result is a possibly empty unconstrained 1-dimensional array containing the elements of the collection that test True when compared to the value designated by the right operand, and $matches is not set. The right operand may designate a string that contains regular expressions (§3.16), in which case, it is referred to as a pattern. These operators have two variants (§7.8).

These operators support submatches (§7.8.4.6).

Examples:

"Hello" -match ".l"                    # True, $matches key/value is 0/"el"
"Hello" -match '\^h.*o$'               # True, $matches key/value is
0/"Hello"
"Hello" -cmatch '\^h.*o$'              # False, $matches not set
"abc\^ef" -match ".\\\^e"              # True, $matches key/value is 0/"c\^e"

"abc" -notmatch "[A-Za-z]"             # False
"abc" -match "[\^A-Za-z]"              # False
"He" -match "h[aeiou]."                # False, need at least 3 characters
"abc","abbcde","abcgh" -match "abc.*"  # Length is 2, values "abc", "abcgh"

7.8.4.3 The -replace operator

Syntax:

binary-replace-operator: one of
    dash replace    dash creplace   dash ireplace

dash:
    - (U+002D)
    EnDash character (U+2013)
    EmDash character (U+2014)
    Horizontal bar character (U+2015)

Description:

The -replace operator allows text replacement in one or more strings designated by the left operand using the values designated by the right operand. This operator has two variants (§7.8). The right operand has one of the following forms:

  • The string to be located, which may contain regular expressions (§3.16). In this case, the replacement string is implicitly "".
  • An array of 2 objects containing the string to be located, followed by the replacement string.

If the left operand designates a string, the result has type string. If the left operand designates a 1-dimensional array of string, the result is an unconstrained 1-dimensional array, whose length is the same as for left operand's array, containing the input strings after replacement has completed.

This operator supports submatches (§7.8.4.6).

Examples:

"Analogous","an apple" -replace "a","*"      # "*n*logous","*n *pple"
"Analogous" -creplace "[aeiou]","?"          # "An?l?g??s"
"Analogous","an apple" -replace '\^a',"%%A"  # "%%Analogous","%%An apple"
"Analogous" -replace "[aeiou]",'$&$&'        # "AAnaaloogoouus"

7.8.4.4 The binary -join operator

Syntax:

binary-join-operator: one of
    dash join

dash:
    - (U+002D)
    EnDash character (U+2013)
    EmDash character (U+2014)
    Horizontal bar character (U+2015)

Description:

The binary -join operator produces a string that is the concatenation of the value of one or more objects designated by the left operand after having been converted to string (§6.7), if necessary. The string designated by the right operand is used to separate the (possibly empty) values in the resulting string.

The left operand can be a scalar value or a collection.

Examples:

(10, 20, 30) -join "\|"    # result is "10\|20\|30"
12345 -join ","            # result is "12345", no separator needed
($null,$null) -join "<->"  # result is "<->", two zero-length values

7.8.4.5 The binary -split operator

Syntax:

binary-split-operator: one of
    dash split      dash csplit     dash isplit

dash:
    - (U+002D)
    EnDash character (U+2013)
    EmDash character (U+2014)
    Horizontal bar character (U+2015)

Description:

The binary -split operator splits one or more strings designated by the left operand, returning their subparts in a constrained 1-dimensional array of string. This operator has two variants (§7.8). The left operand can designate a scalar value or an array of strings. The right operand has one of the following forms:

  • A delimiter string
  • An array of 2 objects containing a delimiter string followed by a numeric split count
  • An array of 3 objects containing a delimiter string, a numeric split count, and an options string
  • A script block
  • An array of 2 objects containing a script block followed by a numeric split count

The delimiter string may contain regular expressions (§3.16). It is used to locate subparts with the input strings. The delimiter is not included in the resulting strings. If the left operand designates an empty string, that results in an empty string element. If the delimiter string is an empty string, it is found at every character position in the input strings.

By default, all subparts of the input strings are placed into the result as separate elements; however, the split count can be used to modify this behavior. If that count is negative, zero, or greater than or equal to the number of subparts in an input string, each subpart goes into a separate element. If that count is less than the number of subparts in the input string, there are count elements in the result, with the final element containing all of the subparts beyond the first count - 1 subparts.

An options string contains zero or more option names with each adjacent pair separated by a comma. Leading, trailing, and embedded whitespace is ignored. Option names may be in any order and are case-sensitive.

If an options string contains the option name SimpleMatch, it may also contain the option name IgnoreCase. If an options string contains the option name RegexMatch or it does not contain either RegexMatch or SimpleMatch, it may contain any option name except SimpleMatch. However, it must not contain both Multiline and Singleline.

Here is the set of option names:

Option Description
CultureInvariant Ignores cultural differences in language when evaluating the delimiter.
ExplicitCapture Ignores non-named match groups so that only explicit capture groups are returned in the result list.
IgnoreCase Force case-insensitive matching, even if -csplit is used.
IgnorePatternWhitespace Ignores unescaped whitespace and comments marked with the number sign (#).
Multiline This mode recognizes the start and end of lines and strings. The default mode is Singleline.
RegexMatch Use regular expression matching to evaluate the delimiter. This is the default.
SimpleMatch Use simple string comparison when evaluating the delimiter.
Singleline This mode recognizes only the start and end of strings. It is the default mode.

The script block (§7.1.8) specifies the rules for determining the delimiter, and must evaluate to type bool.

Examples:

"one,forty two,," -split ","              # 5 strings: "one" "forty two" "" ""

"abc","de" -split ""                      # 9 strings: "" "a" "b" "c" "" "" "d" "e" ""

"ab,cd","1,5,7,8" -split ",", 2           # 4 strings: "ab" "cd" "1" "5,7,8"

"10X20x30" -csplit "X", 0, "SimpleMatch"  # 2 strings: "10" "20x30"

"analogous" -split "[AEIOU]", 0, "RegexMatch, IgnoreCase"
                                          # 6 strings: "" "n" "l" "g" "" "s"

"analogous" -split { $_ -eq "a" -or $_ -eq "o" }, 4
                                          # 4 strings: "" "n" "l" "gous"

7.8.4.6 Submatches

The pattern being matched by -match, -notmatch, and -replace may contain subparts (called submatches) delimited by parentheses. Consider the following example:

"red" -match "red"

The result is $true and key 0 of $matches contains "red", that part of the string designated by the left operand that exactly matched the pattern designated by the right operand.

In the following example, the whole pattern is a submatch:

"red" -match "(red)"

As before, key 0 contains "red"; however, key 1 also contains "red", which is that part of the string designated by the left operand that exactly matched the submatch.

Consider the following, more complex, pattern:

"red" -match "((r)e)(d)"

This pattern allows submatches of "re", "r", "d", or "red".

Again, key 0 contains "red". Key 1 contains "re", key 2 contains "r", and key 3 contains "d". The key/value pairs are in matching order from left-to-right in the pattern, with longer string matches preceding shorter ones.

In the case of -replace, the replacement text can access the submatches via names of the form $n, where the first match is $1, the second is $3, and so on. For example,

"Monday morning" -replace '(Monday|Tuesday) (morning|afternoon|evening)','the $2 of $1'

The resulting string is "the morning of Monday".

Instead of having keys in $matches be zero-based indexes, submatches can be named using the form ?<*name*>. For example, "((r)e)(d)" can be written with three named submatches, m1, m2, and m3, as follows: "(?<m1>(?<m2>r)e)(?<m3>d)".

7.8.5 Shift operators

Syntax:

shift-operator: one of
    dash shl
    dash shr

dash:
    - (U+002D)
    EnDash character (U+2013)
    EmDash character (U+2014)
    Horizontal bar character (U+2015)

Description:

The shift left (-shl) operator and shift right (-shr) operator convert the value designed by the left operand to an integer type and the value designated by the right operand to int, if necessary, using the usual arithmetic conversions (§6.15).

The shift left operator shifts the left operand left by a number of bits computed as described below. The low-order empty bit positions are set to zero.

The shift right operator shifts the left operand right by a number of bits computed as described below. The low-order bits of the left operand are discarded, the remaining bits shifted right. When the left operand is a signed value, the high-order empty bit positions are set to zero if the left operand is non-negative and set to one if the left operand is negative. When the left operand is an unsigned value, the high-order empty bit positions are set to zero.

When the left operand has type int, the shift count is given by the low-order five bits of the right operand. When the right operand has type long, the shift count is given by the low-order six bits of the right operand.

Examples:

0x0408 -shl 1             # int with value 0x0810
0x0408 -shr 3             # int with value 0x0081
0x100000000 -shr 0xfff81  # long with value 0x80000000

7.9 Bitwise operators

Syntax:

bitwise-expression:
    unary-expression -band new-lines~opt~ unary-expression
    unary-expression -bor new-lines~opt~ unary-expression
    unary-expression -bxor new-lines~opt~ unary-expression

Description:

The bitwise AND operator -band, the bitwise OR operator -bor, and the bitwise XOR operator -bxor convert the values designated by their operands to integer types, if necessary, using the usual arithmetic conversions (§6.15). After conversion, if both values have type int that is the type of the result. Otherwise, if both values have type long, that is the type of the result. If one value has type int and the other has type long, the type of the result is long. Otherwise, the expression is ill formed. The result is the bitwise AND, bitwise OR, or bitwise XOR, respectively, of the possibly converted operand values.

These operators are left associative. They are commutative if neither operand contains a side effect.

Examples:

0x0F0F -band 0xFE    # int with value 0xE
0x0F0F -band 0xFEL   # long with value 0xE
0x0F0F -band 14.6    # long with value 0xF

0x0F0F -bor 0xFE     # int with value 0xFFF
0x0F0F -bor 0xFEL    # long with value 0xFFF
0x0F0F -bor 14.40D   # long with value 0xF0F

0x0F0F -bxor 0xFE    # int with value 0xFF1
0x0F0F -bxor 0xFEL   # long with value 0xFF1
0x0F0F -bxor 14.40D  # long with value 0xF01
0x0F0F -bxor 14.6    # long with value 0xF00

7.10 Logical operators

Syntax:

logical-expression:
    unary-expression -and new-lines~opt~ unary-expression
    unary-expression -or new-lines~opt~ unary-expression
    unary-expression -xor new-lines~opt~ unary-expression

Description:

The logical AND operator -and converts the values designated by its operands to bool, if necessary (§6.2). The result is the logical AND of the possibly converted operand values, and has type bool. If the left operand evaluates to False the right operand is not evaluated.

The logical OR operator -or converts the values designated by its operands to bool, if necessary (§6.2). The result is the logical OR of the possibly converted operand values, and has type bool. If the left operand evaluates to True the right operand is not evaluated.

The logical XOR operator -xor converts the values designated by its operands to bool (§6.2). The result is the logical XOR of the possibly converted operand values, and has type bool.

These operators are left associative.

Examples:

$j = 10
$k = 20
($j -gt 5) -and (++$k -lt 15)   # True -and False -> False
($j -gt 5) -and ($k -le 21)     # True -and True -> True
($j++ -gt 5) -and ($j -le 10)   # True -and False -> False
($j -eq 5) -and (++$k -gt 15)   # False -and True -> False

$j = 10
$k = 20
($j++ -gt 5) -or (++$k -lt 15)  # True -or False -> True
($j -eq 10) -or ($k -gt 15)     # False -or True -> True
($j -eq 10) -or (++$k -le 20)   # False -or False -> False

$j = 10
$k = 20
($j++ -gt 5) -xor (++$k -lt 15) # True -xor False -> True
($j -eq 10) -xor ($k -gt 15)    # False -xor True -> True
($j -gt 10) -xor (++$k -le 25)  # True -xor True -> False

7.11 Assignment operators

Syntax:

assignment-expression:
    expression assignment-operator statement

assignment-operator: *one of
    =   dash =   +=   *=   /=   %=

Description:

An assignment operator stores a value in the writable location designated by expression. For a discussion of assignment-operator = see §7.11.1. For a discussion of all other assignment operators see §7.11.2.

An assignment expression has the value designated by expression after the assignment has taken place; however, that assignment expression does not itself designate a writable location. If expression is type-constrained (§5.3), the type used in that constraint is the type of the result; otherwise, the type of the result is the type after the usual arithmetic conversions (§6.15) have been applied.

This operator is right associative.

7.11.1 Simple assignment

Description:

In simple assignment (=), the value designated by statement replaces the value stored in the writable location designated by expression. However, if expression designates a non-existent key in a Hashtable, that key is added to the Hashtable with an associated value of the value designated by statement.

As shown by the grammar, expression may designate a comma-separated list of writable locations. This is known as multiple assignment. statement designates a list of one or more comma-separated values. The commas in either operand list are part of the multiple-assignment syntax and do not represent the binary comma operator. Values are taken from the list designated by statement, in lexical order, and stored in the corresponding writable location designated by expression. If the list designated by statement has fewer values than there are expression writable locations, the excess locations take on the value $null. If the list designated by statement has more values than there are expression writable locations, all but the right-most expression location take on the corresponding statement value and the right-most expression location becomes an unconstrained 1-dimensional array with all the remaining statement values as elements.

For statements that have values (§8.1.2), statement can be a statement.

Examples:

$a = 20; $b = $a + 12L             # $b has type long, value 22
$hypot = [Math]::Sqrt(3*3 + 4*4)   # type double, value 5
$a = $b = $c = 10.20D              # all have type decimal, value 10.20
$a = (10,20,30),(1,2)              # type [object[]], Length 2
[int]$x = 10.6                     # type int, value 11
[long]$x = "0xabc"                 # type long, value 0xabc
$a = [float]                       # value type literal [float]
$i,$j,$k = 10,"red",$true          # $i is 10, $j is "red", $k is True
$i,$j = 10,"red",$true             # $i is 10, $j is [object[]], Length 2
$i,$j = (10,"red"),$true           # $i is [object[]], Length 2, $j is True
$i,$j,$k = 10                      # $i is 10, $j is $null, $k is $null

$h = @{}
[int] $h.Lower, [int] $h.Upper = -split "10 100"

$h1 = @{ FirstName = "James"; LastName = "Anderson"; IDNum = 123 }
$h1.Dept = "Finance"               # adds element Finance
$h1["City"] = "New York"           # adds element City

[int]$Variable:v = 123.456         # v takes on the value 123
${E:output.txt} = "a"              # write text to the given file
$Env:MyPath = "x:\data\file.txt"   # define the environment variable
$Function:F = { param ($a, $b) "Hello there, $a, $b" }
F 10 "red"                         # define and invoke a function
function Demo { "Hi there from inside Demo" }
$Alias:A = "Demo"                  # create alias for function Demo
A                                  # invoke function Demo via the alias

7.11.2 Compound assignment

Description:

A compound assignment has the form E1 op= E2, and is equivalent to the simple assignment expression E1 = E1 op (E2) except that in the compound assignment case the expression E1 is evaluated only once. If expression is type-constrained (§5.3), the type used in that constraint is the type of the result; otherwise, the type of the result is determined by op. For *=, see §7.6.1, §7.6.2, §7.6.3; for /=, see §7.6.4; for %=, see §7.6.5; for +=, see §7.7.1, §7.7.2, §7.7.3; for -=, see §7.7.5.

Note

An operand designating an unconstrained value of numeric type may have its type changed by an assignment operator when the result is stored.

Examples:

$a = 1234; $a *= (3 + 2)  # type is int, value is 1234 * (3 + 2)
$b = 10,20,30             # $b[1] has type int, value 20
$b[1] /= 6                # $b[1] has type double, value 3.33...

$i = 0
$b = 10,20,30
$b[++$i] += 2             # side effect evaluated only once

[int]$Variable:v = 10     # v takes on the value 10
$Variable:v -= 3          # 3 is subtracted from v

${E:output.txt} = "a"     # write text to the given file
${E:output.txt} += "b"    # append text to the file giving ab
${E:output.txt} *= 4      # replicate ab 4 times giving abababab

7.12 Redirection operators

Syntax:

pipeline:
    expression redirections~opt~ pipeline-tail~opt~
    command verbatim-command-argument~opt~ pipeline-tail~opt~

redirections:
    redirection
    redirections redirection

redirection:
    merging-redirection-operator
    file-redirection-operator redirected-file-name

redirected-file-name:
    command-argument
    primary-expression

file-redirection-operator: one of
    >   >>   2>   2>>   3>   3>>   4>   4>>
    5>  5>>  6>   6>>   >    >>    <

merging-redirection-operator: one of
    >&1   2>&1   3>&1   4>&1   5>&1   6>&1
    >&2   1>&2   3>&2   4>&2   5>&2   6>&2

Description:

The redirection operator > takes the standard output from the pipeline and redirects it to the location designated by redirected-file-name, overwriting that location's current contents.

The redirection operator >> takes the standard output from the pipeline and redirects it to the location designated by redirected-file-name, appending to that location's current contents, if any. If that location does not exist, it is created.

The redirection operator with the form n> takes the output of stream n from the pipeline and redirects it to the location designated by redirected-file-name, overwriting that location's current contents.

The redirection operator with the form n>> takes the output of stream n from the pipeline and redirects it to the location designated by redirected-file-name, appending to that location's current contents, if any. If that location does not exist, it is created.

The redirection operator with the form m>&n writes output from stream m to the same location as stream n.

The following are the valid streams:

Stream Description
1 Standard output stream
2 Error output stream
3 Warning output stream
4 Verbose output stream
5 Debug output stream
* Standard output, error output, warning output, verbose output, and debug output streams

The redirection operators 1>&2, 6>, 6>> and < are reserved for future use.

If on output the value of redirected-file-name is $null, the output is discarded.

Ordinarily, the value of an expression containing a top-level side effect is not written to the pipeline unless that expression is enclosed in a pair of parentheses. However, if such an expression is the left operand of an operator that redirects standard output, the value is written.

Examples:

$i = 200                       # pipeline gets nothing
$i                             # pipeline gets result
$i > output1.txt               # result redirected to named file
++$i >> output1.txt            # result appended to named file
type file1.txt 2> error1.txt   # error output redirected to named file
type file2.txt 2>> error1.txt  # error output appended to named file
dir -Verbose 4> verbose1.txt   # verbose output redirected to named file

# Send all output to output2.txt
dir -Verbose -Debug -WarningAction Continue *> output2.txt

# error output redirected to named file, verbose output redirected
# to the same location as error output
dir -Verbose 4>&2 2> error2.txt