[jira] [Created] (SPARK-7247) Add Pandas' shift method to the Dataframe API

["Olivier Girardot (JIRA)" <ji...@apache.org>]

Olivier Girardot created SPARK-7247:
---------------------------------------

             Summary: Add Pandas' shift method to the Dataframe API
                 Key: SPARK-7247
                 URL: https://issues.apache.org/jira/browse/SPARK-7247
             Project: Spark
          Issue Type: Wish
          Components: SQL
    Affects Versions: 1.3.1
            Reporter: Olivier Girardot
            Priority: Minor


Spark's DataFrame provide several of the capabilities of Pandas and R Dataframe but in a distributed fashion which makes it almost easy to rewrite Pandas code to Spark.

Almost, but there is a feature that's difficult to workaround right now, that's the shift method : http://pandas.pydata.org/pandas-docs/stable/generated/pandas.DataFrame.shift.html

I'm working with some data scientists that are using this feature a lot in order to check for "rows equality" after a sort by some keys. Example (in pandas) :

{code}
df['delta'] = (df.START_DATE.shift(-1) - df.END_DATE).astype('timedelta64[D]')
{code}

I think this would be troublesome to add, I don't even know if this change would be do-able. But as a user, it would be useful for me.

Olivier.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org