共用方式為

INVALID_ARRAY_INDEX_IN_ELEMENT_AT error condition

SQLSTATE: 22003

The index <indexValue> is out of bounds. 陣列具有 <arraySize> 元素。 Use try_element_at to tolerate accessing element at invalid index and return NULL instead. 如有必要，請將 <ansiConfig> 設為 「false」 以略過此錯誤。

參數

  • indexValue: The requested index into the array.
  • arraySize：陣列的基數。
  • ansiConfig：要改變 ANSI 模式的組態設定。

解釋

indexValue 超過 element_at（arrayExpr, indexValue）或 elt（arrayExpr, indexValue） 表達式的已定義陣列元素的範圍。

值必須介於 -arraySize 和 arraySize之間，不包括 0。

Mitigation

此錯誤的緩解取決於原因：

  • 陣列的基數是否小於預期？

    修正輸入陣列並重新執行查詢。

  • indexValue 是否計算錯誤？

    調整 indexValue 並重新執行查詢。

  • 您是否預期會針對索引基數以外的項目傳回 NULL 值？

    如果您可以變更表達式，請使用 try_element_at(arrayExpr, indexValue) 來容許超出範圍的存取。

    If you cannot change the expression, as a last resort, temporarily set the ansiConfig to false to tolerate references out of bound.

例子

-- An INVALID_ARRAY_INDEX_IN_ELEMENT_AT error because of mismatched indexing
> SELECT element_at(array('a', 'b', 'c'), index) FROM VALUES(1), (4) AS T(index);
  [INVALID_ARRAY_INDEX_IN_ELEMENT_AT] The index 4 is out of bounds. The array has 3 elements. If necessary set "ANSI_MODE" to false to bypass this error.

-- Increase the aray size to cover the index
> SELECT element_at(array('a', 'b', 'c', 'd'), index) FROM VALUES(1), (4) AS T(index);
  a
  d

-- Adjusting the index to match the array
> SELECT element_at(array('a', 'b', 'c'), index) FROM VALUES(1), (3) AS T(index);
  a
  c

-- Tolerating out of bound array index with adjustment to 1-based indexing
> SELECT try_element_at(array('a', 'b', 'c'), index) FROM VALUES(1), (4) AS T(index);
  a
  NULL

-- Tolerating out of bound by setting ansiConfig in Databricks SQL
> SET ANSI_MODE = false;
> SELECT element_at(array('a', 'b', 'c'), index) FROM VALUES(1), (4) AS T(index);
  a
  NULL
> SET ANSI_MODE = true;

-- Tolerating out of bound by setting ansiConfig in Databricks Runtime
> SET spark.sql.ansi.enabled = false;
> SELECT element_at(array('a', 'b', 'c'), index) FROM VALUES(1), (4) AS T(index);
  a
  NULL
> SET spark.sql.ansi.enabled = true;