將結構陣列分解成數據表。
此函式會接收包含結構數陣列的輸入欄位，並回傳一個新欄位，將陣列中的每個結構體拆成獨立的列。
語法
from pyspark.sql import functions as sf
sf.inline(col)
參數
| 參數 | 類型 | Description |
|---|---|---|
col |
pyspark.sql.Column 或欄位名稱 |
輸入欄位的數值會爆炸。 |
退貨
pyspark.sql.Column：產生器表達式，並取得內聯爆炸結果。
範例
範例 1：使用單一結構陣列欄位的內嵌
import pyspark.sql.functions as sf
df = spark.sql('SELECT ARRAY(NAMED_STRUCT("a",1,"b",2), NAMED_STRUCT("a",3,"b",4)) AS a')
df.select('*', sf.inline(df.a)).show()
+----------------+---+---+
| a| a| b|
+----------------+---+---+
|[{1, 2}, {3, 4}]| 1| 2|
|[{1, 2}, {3, 4}]| 3| 4|
+----------------+---+---+
範例 2：使用欄位名稱的內嵌
import pyspark.sql.functions as sf
df = spark.sql('SELECT ARRAY(NAMED_STRUCT("a",1,"b",2), NAMED_STRUCT("a",3,"b",4)) AS a')
df.select('*', sf.inline('a')).show()
+----------------+---+---+
| a| a| b|
+----------------+---+---+
|[{1, 2}, {3, 4}]| 1| 2|
|[{1, 2}, {3, 4}]| 3| 4|
+----------------+---+---+
範例三：使用內嵌名稱
import pyspark.sql.functions as sf
df = spark.sql('SELECT ARRAY(NAMED_STRUCT("a",1,"b",2), NAMED_STRUCT("a",3,"b",4)) AS a')
df.select('*', sf.inline('a').alias("c1", "c2")).show()
+----------------+---+---+
| a| c1| c2|
+----------------+---+---+
|[{1, 2}, {3, 4}]| 1| 2|
|[{1, 2}, {3, 4}]| 3| 4|
+----------------+---+---+
範例 4：使用多重結構陣列欄位的內嵌
import pyspark.sql.functions as sf
df = spark.sql('SELECT ARRAY(NAMED_STRUCT("a",1,"b",2), NAMED_STRUCT("a",3,"b",4)) AS a1, ARRAY(NAMED_STRUCT("c",5,"d",6), NAMED_STRUCT("c",7,"d",8)) AS a2')
df.select(
'*', sf.inline('a1')
).select('*', sf.inline('a2')).show()
+----------------+----------------+---+---+---+---+
| a1| a2| a| b| c| d|
+----------------+----------------+---+---+---+---+
|[{1, 2}, {3, 4}]|[{5, 6}, {7, 8}]| 1| 2| 5| 6|
|[{1, 2}, {3, 4}]|[{5, 6}, {7, 8}]| 1| 2| 7| 8|
|[{1, 2}, {3, 4}]|[{5, 6}, {7, 8}]| 3| 4| 5| 6|
|[{1, 2}, {3, 4}]|[{5, 6}, {7, 8}]| 3| 4| 7| 8|
+----------------+----------------+---+---+---+---+
範例 5：使用內嵌結構陣列欄位
import pyspark.sql.functions as sf
df = spark.sql('SELECT NAMED_STRUCT("a",1,"b",2,"c",ARRAY(NAMED_STRUCT("c",3,"d",4), NAMED_STRUCT("c",5,"d",6))) AS s')
df.select('*', sf.inline('s.c')).show(truncate=False)
+------------------------+---+---+
|s |c |d |
+------------------------+---+---+
|{1, 2, [{3, 4}, {5, 6}]}|3 |4 |
|{1, 2, [{3, 4}, {5, 6}]}|5 |6 |
+------------------------+---+---+
範例 6：使用內聯，欄位包含：包含 null、empty 陣列和 null 的陣列
from pyspark.sql import functions as sf
df = spark.sql('SELECT * FROM VALUES (1,ARRAY(NAMED_STRUCT("a",1,"b",2), NULL, NAMED_STRUCT("a",3,"b",4))), (2,ARRAY()), (3,NULL) AS t(i,s)')
df.show(truncate=False)
+---+----------------------+
|i |s |
+---+----------------------+
|1 |[{1, 2}, NULL, {3, 4}]|
|2 |[] |
|3 |NULL |
+---+----------------------+
df.select('*', sf.inline('s')).show(truncate=False)
+---+----------------------+----+----+
|i |s |a |b |
+---+----------------------+----+----+
|1 |[{1, 2}, NULL, {3, 4}]|1 |2 |
|1 |[{1, 2}, NULL, {3, 4}]|NULL|NULL|
|1 |[{1, 2}, NULL, {3, 4}]|3 |4 |
+---+----------------------+----+----+