Spark dataFrame將列數據類型從字符串轉換為日期

Question

我有以下數據，有架構

scala> df2.printSchema()
root
 |-- RowID: integer (nullable = true)
 |-- Order Date: string (nullable = true)

scala> df2.show(5)
+-----+----------+
|RowID|Order Date|
+-----+----------+
|    1|   4/10/15|
|   49|   4/10/15|
|   50|   4/10/15|
|   80|   4/10/15|
|   85|   4/10/15|
+-----+----------+

我想將“訂購日期”字符串列轉換為日期數據類型，然后嘗試以下操作，但沒有運氣，有人可以建議一種更好的方法嗎？

scala> df2.select(df2.col("RowID"), df2.col("Order Date"), date_format(df2.col("Order Date"), "M/dd/yy")).show(5)
+-----+----------+-------------------------------+
|RowID|Order Date|date_format(Order Date,M/dd/yy)|
+-----+----------+-------------------------------+
|    1|   4/10/15|                           null|
|   49|   4/10/15|                           null|
|   50|   4/10/15|                           null|
|   80|   4/10/15|                           null|
|   85|   4/10/15|                           null|
+-----+----------+-------------------------------+

Answer 1

設法轉換為unix epoch時間戳，我認為從這里開始很簡單

scala> df.select(df.col("RowID"), df.col("Order Date"), unix_timestamp(df.col("Order Date"), "M/d/yy")).show(5)
+-----+----------+--------------------------------+
|RowID|Order Date|unixtimestamp(Order Date,M/d/yy)|
+-----+----------+--------------------------------+
|    1|   4/10/15|                      1428604200|
|   49|   4/10/15|                      1428604200|
|   50|   4/10/15|                      1428604200|
|   80|   4/10/15|                      1428604200|
|   85|   4/10/15|                      1428604200|
+-----+----------+--------------------------------+