Try to_json on the vector column. That should do it.
--
Sent from: http://apache-spark-user-list.1001560.n3.nabble.com/
-
To unsubscribe e-mail: user-unsubscr...@spark.apache.org
Hi,
I was hoping that there is a casting vector into String method (instead of
writing my UDF), so that it can then be serialized it into csv/text file.
Best regards,
Mina
On Tue, Feb 20, 2018 at 6:52 PM, vermanurag
wrote:
> If your dataframe has columns types
Hi Snehasish,
Unfortunately, none of the solutions worked.
Regards,
Mina
On Tue, Feb 20, 2018 at 5:12 PM, SNEHASISH DUTTA
wrote:
> Hi Mina,
>
> Even text won't work you may try this df.coalesce(1).write.option("h
>
If your dataframe has columns types like vector then you cannot save as csv/
text as there are no direct equivalent supported by flat formats like csv/
text. You may need to convert the column type appropriately (eg. convert the
incompatible column to StringType before saving the output as csv.
Hi Mina,
Even text won't work you may try this df.coalesce(1).write.option("h
eader","true").mode("overwrite").save("output",format=text)
Else convert to an rdd and use saveAsTextFile
Regards,
Snehasish
On Wed, Feb 21, 2018 at 3:38 AM, SNEHASISH DUTTA
wrote:
> Hi
Hi Mina,
This might work then
df.coalesce(1).write.option("header","true").mode("overwrite
").text("output")
Regards,
Snehasish
On Wed, Feb 21, 2018 at 3:21 AM, Mina Aslani wrote:
> Hi Snehasish,
>
> Using df.coalesce(1).write.option("header","true").mode("overwrite
>
Hi Snehasish,
Using df.coalesce(1).write.option("header","true").mode("overwrite
").csv("output") throws
java.lang.UnsupportedOperationException: CSV data source does not support
struct<...> data type.
Regards,
Mina
On Tue, Feb 20, 2018 at 4:36 PM, SNEHASISH DUTTA
Hi Mina,
This might help
df.coalesce(1).write.option("header","true").mode("overwrite").csv("output")
Regards,
Snehasish
On Wed, Feb 21, 2018 at 1:53 AM, Mina Aslani wrote:
> Hi,
>
> I would like to serialize a dataframe with vector values into a text/csv
> in pyspark.
>