Case sensitive mapping Hive about elasticsearch-hadoop

[marcelopaesrech]

Apparently the mapping from ES to Hive considers case sensitive. But when receives data from ES the plugin just gets the name field from Hive (in this case countryid - in lower case), and try to get from the map returned from ES.

I made a change in this parser, but I didn't test if when the flow is HIVE to ES if its work.

[costin]

Mapping is currently work-in-progress and touches on some of the other issues you reported. It's too early to point to a solution - most likely there will be an option to do field aliasing (i.e. point how the fields in ES map to the fields in Hive/Pig/etc...).

What is your mapping in ES vs Hive? Can't you get around it (for now) by using an intermediate Hive table and then copying things over?

[costin]

@marcelopaesrech can you please try 1.3 M1 or the master? This has been fixed some time ago. The issue was caused by the fact that Hive is case insensitive while ES is which lead to some issues in the way the columns were returned. Additionally, there's also support for mapping (useful when dealing with characters not supported by Hive or Pig) through es.mapping.names
Thanks!