Spark – Add Hours, Minutes, and Seconds to Timestamp

Problem: How to add or subtract hours, minutes, and seconds to/from Spark DataFrame date or timestamp columns?

Solution: Spark SQL has no functions that add/subtract time unit hours, minutes, and seconds to or from a Timestamp column, however, SQL defines Interval to do it.

Refer to Spark SQL Date and Timestamp Functions for all Date & Time functions.

1. Using Spark SQL Interval

Since Spark doesn’t have any functions to add units to the Timestamp, we use INTERVAL to do our job.

// Using Spark SQL Interval
import org.apache.spark.sql.functions._
import spark.sqlContext.implicits._

spark.sql( "select current_timestamp," +
    "cast(current_timestamp as TIMESTAMP) + INTERVAL 2 hours as added_hours," +
    "cast(current_timestamp as TIMESTAMP) + INTERVAL 5 minutes as added_minutes," +
    "cast(current_timestamp as TIMESTAMP) + INTERVAL 55 seconds as added_seconds"
  ).show(false)

Before we apply INTERVAL, first you need to convert timestamp column from string to TimestampType using cast.

Yields below output:

// Output:
+----------------------+----------------------+----------------------+----------------------+
|current_timestamp()   |added_hours           |added_minutes         |added_seconds         |
+----------------------+----------------------+----------------------+----------------------+
|2019-11-17 05:54:22.87|2019-11-17 07:54:22.87|2019-11-17 05:59:22.87|2019-11-17 05:55:17.87|
+----------------------+----------------------+----------------------+----------------------+

The above example directly uses spark select without DataFrame, below example demonstrates how to do the same when we have a timestamp as a DataFrame column

import org.apache.spark.sql.functions._
val df = Seq(("2019-07-01 12:01:19.101"),
    ("2019-06-24 12:01:19.222"),
    ("2019-11-16 16:44:55.406"),
    ("2019-11-16 16:50:59.406")).toDF("input_timestamp")

df.createOrReplaceTempView("AddTimeExample")

val df2 = spark.sql("select input_timestamp, " +
    "cast(input_timestamp as TIMESTAMP) + INTERVAL 2 hours as added_hours," +
    "cast(input_timestamp as TIMESTAMP) + INTERVAL 5 minutes as added_minutes," +
    "cast(input_timestamp as TIMESTAMP) + INTERVAL 55 seconds as added_seconds from AddTimeExample"
    )
df2.show(false)

Here, first, we create a temporary table using createOrReplaceTempView() and then use this on SQL select.

+-----------------------+-----------------------+-----------------------+-----------------------+
|input_timestamp        |added_hours            |added_minutes          |added_seconds          |
+-----------------------+-----------------------+-----------------------+-----------------------+
|2019-07-01 12:01:19.101|2019-07-01 14:01:19.101|2019-07-01 12:06:19.101|2019-07-01 12:02:14.101|
|2019-06-24 12:01:19.222|2019-06-24 14:01:19.222|2019-06-24 12:06:19.222|2019-06-24 12:02:14.222|
|2019-11-16 16:44:55.406|2019-11-16 18:44:55.406|2019-11-16 16:49:55.406|2019-11-16 16:45:50.406|
|2019-11-16 16:50:59.406|2019-11-16 18:50:59.406|2019-11-16 16:55:59.406|2019-11-16 16:51:54.406|
+-----------------------+-----------------------+-----------------------+-----------------------+

Let’s see the same example without creating a temporary table but use <a href="https://sparkbyexamples.com/spark/spark-dataframe-withcolumn/">withColumn()</a> and expr() function

import org.apache.spark.sql.functions._
df.withColumn("added_hours",col("input_timestamp") + expr("INTERVAL 2 HOURS"))
  .withColumn("added_minutes",col("input_timestamp") + expr("INTERVAL 2 minutes"))
  .withColumn("added_seconds",col("input_timestamp") + expr("INTERVAL 2 seconds"))
  .show(false)

2. Spark Subtract Hours, Minutes, and Seconds from Timestamp

By just using – instead of +, you can subtract an hour, minute, and second from Timestamp.

// Spark Subtract Hours, Minutes, and Seconds from Timestamp 
import org.apache.spark.sql.functions._
df.withColumn("sub_hours",col("input_timestamp") - expr("INTERVAL 2 HOURS"))
  .withColumn("sub_minutes",col("input_timestamp") - expr("INTERVAL 2 minutes"))
  .withColumn("sub_seconds",col("input_timestamp") - expr("INTERVAL 2 seconds"))
  .show(false)

3. Complete Example

Below is a complete example of how to add or subtract hours, minutes, and seconds from the DataFrame Timestamp column. This example is also available at Spark Examples Git Hub project

// Complete Example 
package com.sparkbyexamples.spark.dataframe.functions.datetime

import org.apache.spark.sql.SparkSession

object AddTime extends App {

  val spark:SparkSession = SparkSession.builder()
    .master("local")
    .appName("SparkByExamples.com")
    .getOrCreate()
  spark.sparkContext.setLogLevel("ERROR")

  import spark.sqlContext.implicits._

  spark.sql( "select current_timestamp," +
    "cast(current_timestamp as TIMESTAMP) + INTERVAL 2 hours as added_hours," +
    "cast(current_timestamp as TIMESTAMP) + INTERVAL 5 minutes as added_minutes," +
    "cast(current_timestamp as TIMESTAMP) + INTERVAL 55 seconds as added_seconds"
  ).show(false)


  val df = Seq(("2019-07-01 12:01:19.101"),
    ("2019-06-24 12:01:19.222"),
    ("2019-11-16 16:44:55.406"),
    ("2019-11-16 16:50:59.406")).toDF("input_timestamp")


  df.createOrReplaceTempView("AddTimeExample")

  val df2 = spark.sql("select input_timestamp, " +
    "cast(input_timestamp as TIMESTAMP) + INTERVAL 2 hours as added_hours," +
    "cast(input_timestamp as TIMESTAMP) + INTERVAL 5 minutes as added_minutes," +
    "cast(input_timestamp as TIMESTAMP) + INTERVAL 55 seconds as added_seconds from AddTimeExample"
    )
  df2.show(false)

  // Alternate way
  df.withColumn("added_hours",col("input_timestamp") + expr("INTERVAL 2 HOURS"))
    .withColumn("added_minutes",col("input_timestamp") + expr("INTERVAL 2 minutes"))
    .withColumn("added_seconds",col("input_timestamp") + expr("INTERVAL 2 seconds"))
    .show(false)
}

Happy Learning!!

Related Articles

• Spark SQL – Add Day, Month, and Year to Date
• Spark Timestamp – Extract hour, minute and second
• Spark SQL – Working with Unix Timestamp
• Spark date_format() – Convert Timestamp to String
• Spark to_timestamp() – Convert String to Timestamp Type
• Spark Add Constant Column to DataFrame
• Add Multiple Jars to Spark Submit Classpath?
• Spark – Add New Column & Multiple Columns to DataFrame
• Cannot call methods on a stopped SparkContext in Spark
• Tune Spark Executor Number, Cores, and Memory
• Calculate difference between two dates in days, months and years