Re: How to load customized map data schema

2012-02-07 Thread Daniel Dai
data = load '/tmp/titan_mixi_prod.10' using he.HEStorage() as (uid:long, payload:map[], ts:long, type:int); Daniel 2012/2/6 Haitao Yao : > I've wrote my own loader and here's the error: > grunt>data = load '/tmp/titan_mixi_prod.10' using he.HEStorage() as > (uid:long, payload:map, ts:long, type:

Re: How to load customized map data schema

2012-02-06 Thread Haitao Yao
I've wrote my own loader and here's the error: grunt>data = load '/tmp/titan_mixi_prod.10' using he.HEStorage() as (uid:long, payload:map, ts:long, type:int); ERROR 1200: mismatched input ',' expecting LEFT_BRACKET Failed to parse: mismatched input ',' expecting LEFT_BRACKET at org

Re: How to load customized map data schema

2012-02-06 Thread Daniel Dai
Currently PigStorage only takes: [key#value,key#value] You will have to preprocess it before using PigStorage. Sure you can also write your own loader. Daniel On Sun, Feb 5, 2012 at 11:12 PM, Haitao Yao wrote: > Hi, all >        out data format for map is Key:Value|Key:Value , how can I load the

How to load customized map data schema

2012-02-05 Thread Haitao Yao
Hi, all out data format for map is Key:Value|Key:Value , how can I load the data into map type? Can pig define the map delimiter like hive? thanks.