Versions Compared

Old Version 1

changes.mady.by.user Atanu Mishra

Saved on

compared with

New Version 2

changes.mady.by.user Suresh Subbiah

Saved on

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

...

Data SourceINSERTUPSERTUPSERT USING LOADLOAD
Single row at a time with the VALUES clause<TBD>. Slowest approach.<TBD><TBD>No
List of rows with the VALUES clause.<TBD>. Speed comparable to rowset INSERT. List Single row version of CheckAndPut called.<TBD>. Speed comparable to rowset UPSERT. List version of Put called.<TBD>. Speed comparable to rowset UPSERT USING LOAD. List version of Put called.No
List of rows with ODBC/JDBC Rowset<TBD><TBD>~ 2.5 GB/hrNo
Native HBase tables<TBD>. Speed expected to be slower than a Trafodion table as the source for INSERT since the compiler has no stats and data will need to be encoded.<TBD>. Speed expected to be slower than a Trafodion table as the source for UPSERT since the compiler has no stats and data will need to be encoded.<TBD>. Speed expected to be slower than a Trafodion table as the source for UPSERT USING LOAD since the compiler has no stats and data will need to be encoded.<TBD>. Speed expected to be slower than a Trafodion table as the source for LOAD since the compiler has no stats and data will need to be encoded.
Trafodion tables<TBD>. Speed expected to be slower than a Hive table as the source for INSERT since a Trafodion scan is slower than a Hive scan.<TBD>. Speed expected to be slower than a Hive table as the source for UPSERT since a Trafodion scan is slower than a Hive scan.<TBD>. Speed expected to be slower than a Hive table as the source for UPSERT USING LOAD since a Trafodion scan is slower than a Hive scan.<TBD>. Speed expected to be slower than a Hive table as the source for LOAD since a Trafodion scan is slower than a Hive scan.
Hive tables<TBD>. Fastest source for INSERT.<TBD>. Fastest source for UPSERT.TBD>. Fastest source for UPSERT USING LOAD.Fastest, ~100 GB/hr.
UDF connected to other data sources.Speed depends on the data source and UDF. Expected to be slower than a Hive table as the source for INSERTSpeed depends on the data source and UDF.Speed depends on the data source and UDF.Speed depends on the data source and UDF.

...