When importing a file into Greenplum,one lines fails,and the whole file is not imported successfully.Is there a way can skip the wrong line and import other data into Greenplum successfully?
Here are my SQL execution and error messages:
copy cjh_test from '/gp_wkspace/outputs/base_tables/error_data_test.csv' using delimiters ',';
ERROR: invalid input syntax for integer: "FE00F760B39BD3756BCFF30000000600" CONTEXT: COPY cjh_test, line 81, column local_city: "FE00F760B39BD3756BCFF30000000600"
Greenplum has an extension to the COPY command that lets you log errors and set up a certain amount of errors that can occur that won't stop the load. Here is an example from the documentation for the COPY command:
COPY sales FROM '/home/usr1/sql/sales_data' LOG ERRORS
SEGMENT REJECT LIMIT 10 ROWS;
That tells COPY that 10 bad rows can be ignored without stopping the load. The reject limit can be # of rows or a percentage of the load file. You can check the full syntax in psql with: \h copy
If you are loading a very large file into Greenplum, I would suggest looking at gpload or gpfdist (which also support the segment reject limit syntax). COPY is single threaded through the master server where gpload/gpfdist load the data in parallel to all segments. COPY will be faster for smaller load files and the others will be faster for millions of rows in a load file(s).