Koalas: DataFrame.apply(func, axis=1)

[patryk-oleniuk]

I'm missing the access of the whole row in the DataFrame.apply function, like

df = pd.DataFrame({"date":pd.to_datetime(['2012-01-01','2012-01-02']),
              "hour": [5, 5], 
              })

print(df)

df['datetime'] = df.apply(lambda x: x['date'].replace(hour=x['hour']), axis=1)

print(df)

Prints out:

        date  hour
0 2012-01-01     5
1 2012-01-02     5

        date  hour            datetime
0 2012-01-01     5 2012-01-01 05:00:00
1 2012-01-02     5 2012-01-02 05:00:00

We should implement DataFrame.apply(row_func, axis=1) using the UDFs.

[ueshin]

related to #65

[HyukjinKwon]

I'll start work on this soon. Maybe it will be impossible to support exactly same API but should be able to provide workaround.

[HyukjinKwon]

I made a PR https://github.com/databricks/koalas/pull/1259