spark scala split timestamp column in date column and time column

Question

I have a problem with split of a timestamp column into Date and time columns. First the time doesn't consider 24h format ... Second the date is false and i don't understand why

here is my output

+----------+----------+-------------------+---------+
|      Date| Timestamp|               Time|EventTime|
+----------+----------+-------------------+---------+
|2018-00-30|1540857600|2018-10-30 00:00:00| 12:00:00|
|2018-00-30|1540857610|2018-10-30 00:00:10| 12:00:10|
|2018-00-30|1540857620|2018-10-30 00:00:20| 12:00:20|
|2018-00-30|1540857630|2018-10-30 00:00:30| 12:00:30|
|2018-00-30|1540857640|2018-10-30 00:00:40| 12:00:40|
|2018-00-30|1540857650|2018-10-30 00:00:50| 12:00:50|
|2018-01-30|1540857660|2018-10-30 00:01:00| 12:01:00|
|2018-01-30|1540857670|2018-10-30 00:01:10| 12:01:10|
|2018-01-30|1540857680|2018-10-30 00:01:20| 12:01:20|
|2018-01-30|1540857690|2018-10-30 00:01:30| 12:01:30|
|2018-01-30|1540857700|2018-10-30 00:01:40| 12:01:40|

and my code :

  val df = data_input
    .withColumn("Time", to_timestamp(from_unixtime(col("Timestamp"))))
    .withColumn("Date", date_format(col("Time"), "yyyy-mm-dd"))
    .withColumn("EventTime", date_format(col("Time"), "hh:mm:ss"))

first I convert the unix Timestamp column to Time column, and then i want to split Time..

Thank you in advance

Answer 1

You are using the wrong format codes. Specifically "mm" in your date is for minutes, and "hh" is for 12 hour values. Instead you want "MM" and "HH". Like this:

val df = data_input
    .withColumn("Time", to_timestamp(from_unixtime(col("Timestamp"))))
    .withColumn("Date", date_format(col("Time"), "yyyy-MM-dd"))
    .withColumn("EventTime", date_format(col("Time"), "HH:mm:ss"))

For reference, here are the date format codes that you can use: SimpleDateFormat

Answer 2

You can avoid the confusion with simple casting

import org.apache.spark.sql.functions._

val df = data_input
    .withColumn("Time", $"Timestamp".cast("timestamp"))
    .withColumn("Date", $"Time".cast("date"))
    .withColumn("EventTime", date_format($"Time", "H:m:s"))

+----------+-------------------+----------+---------+
|Timestamp |               Time|      Date|EventTime|
+----------+-------------------+----------+---------+
|1540857600|2018-10-30 00:00:00|2018-10-30|    0:0:0|
|1540857610|2018-10-30 00:00:10|2018-10-30|   0:0:10|
|1540857620|2018-10-30 00:00:20|2018-10-30|   0:0:20|
|1540857630|2018-10-30 00:00:30|2018-10-30|   0:0:30|
|1540857640|2018-10-30 00:00:40|2018-10-30|   0:0:40|
|1540857650|2018-10-30 00:00:50|2018-10-30|   0:0:50|
|1540857660|2018-10-30 00:01:00|2018-10-30|    0:1:0|
|1540857670|2018-10-30 00:01:10|2018-10-30|   0:1:10|
|1540857680|2018-10-30 00:01:20|2018-10-30|   0:1:20|
|1540857690|2018-10-30 00:01:30|2018-10-30|   0:1:30|
|1540857700|2018-10-30 00:01:40|2018-10-30|   0:1:40|
+----------+-------------------+----------+---------+