https://community.cloudera.com/t5/Support-Questions/Best-Practice-JSON-to-Avro-data-type-preserving/m-p/148373#M110899 <P>What are you using for your AVRO Schema? If you specify that the field as timestamp.</P><P><A href="https://avro.apache.org/docs/1.8.0/spec.html#Timestamp+%28millisecond+precision%29" target="_blank">https://avro.apache.org/docs/1.8.0/spec.html#Timestamp+%28millisecond+precision%29</A></P><P><A href="https://cwiki.apache.org/confluence/display/Hive/LanguageManual+Types#LanguageManualTypes-TimestampstimestampTimestamps" target="_blank">https://cwiki.apache.org/confluence/display/Hive/LanguageManual+Types#LanguageManualTypes-TimestampstimestampTimestamps</A></P><P>Can you post an example JSON file, AVRO Schema, data provenance of the run, Hive DDL</P><P>Standard Format "<CODE>YYYY-MM-DD HH:MM:SS.fffffffff</CODE>"</P><P>For the schema, copy the inferred one from a data provenance run and then change the type from string to timestamp and save that schema and use that for the next run.</P><P>If there's empty strings or null in the first 20 records, I am thinking you will get String as a type.</P> Fri, 27 Jan 2017 01:56:09 GMT TimothySpann 2017-01-27T01:56:09Z