Pyspark obtain time attributes from datetime with list comprehension

Question

I have a pyspark dataframe df:

+-------------------+
|      timestamplast|
+-------------------+
|2019-08-01 00:00:00|
|2019-08-01 00:01:09|
|2019-08-01 01:00:20|
|2019-08-03 00:00:27|
+-------------------+

I want to add columns 'year','month','day','hour' to the existing dataframe by list comprehension.

In Pandas this would be done as such:

L = ['year', 'month', 'day', 'hour'] 
date_gen = (getattr(df['timestamplast'].dt, i).rename(i) for i in L) 
df = df.join(pd.concat(date_gen, axis=1)) # concatenate results and join to original dataframe

How would this be done in pyspark?

Answer 1

check the following:

df.selectExpr("*", *[ '{0}(timestamplast) as {0}'.format(c) for c in L]).show()                                    
+-------------------+----+-----+---+----+
|      timestamplast|year|month|day|hour|
+-------------------+----+-----+---+----+
|2019-08-01 00:00:00|2019|    8|  1|   0|
|2019-08-03 00:00:27|2019|    8|  3|   0|
+-------------------+----+-----+---+----+