Reputation: 5706
I am reading Parquet
logs in PIG
. Log schema contains a field FULL
nested under field USER
. Since FULL
is a keyword in Pig I can't load the file. Pig is giving compile time errors. There is no way to escape such field name. It would really adverse to change column name in schema. This is what I am using to load
user_parquet = LOAD 'test' USING org.apache.parquet.pig.ParquetLoader();
user_normalized = FOREACH user_parquet GENERATE ..... USER.FULL as user_full ...
I tried escaping with \, ``, ''
but none of them worked!
Upvotes: 1
Views: 664
Reputation: 5706
Basically I solved the issue by referring field with field number instead of field name:-
user_parquet = LOAD 'test' USING org.apache.parquet.pig.ParquetLoader();
user_normalized = FOREACH user_parquet GENERATE ..... USER.$0 as user_full ...
Pig, as any other language as no way to escape keyword if used as variable name.
Upvotes: 1