Scalding read Timestamp from parquet

678 views Asked by At

Hi I am trying to read a timestamp field (datetime) from A parquet file using a case class (in a typedpipe).

The parquet file is from an impala table.

the parquet-tools lib gives the type as

optional int96

Does anyone know what the field type should be ?

If I use Option[Long]

case class Columns(request_id:Option[String],datetime: Option[Long])
 val input = TypedParquet[Columns](input1)

then it gives the below error

Caused by: org.apache.parquet.io.ParquetDecodingException: The requested schema is not compatible with the file schema. incompatible types: optional int64 datetime != optional int96 datetime

which kind of makes sense.

If I try to use BigInt

case class Columns(request_id:Option[String],datetime: Option[Bigint])
 val input = TypedParquet[Columns](input1)

It does not even compile it gives the below error

Parquetreadsupport - could not find implicit value for parameter

0

There are 0 answers