Uses the gpfdist protocol to create a readable external table, ext_expenses, from all files with the txt extension. The column delimiter is a pipe ( | ) and NULL (' ') is a space.

Access to the external table is single row error isolation mode. Input data formatting errors are captured internally in Greenplum Database with a description of the error. See 在错误日志中查看不正确的行 for information about investigating error rows. You can view the errors, fix the issues, and then reload the rejected data. If the error count on a segment is greater than five (the SEGMENT REJECT LIMIT value), the entire external table operation fails and no rows are processed.

=# CREATE EXTERNAL TABLE ext_expenses ( name text, 
   date date, amount float4, category text, desc1 text ) 
   LOCATION ('gpfdist://etlhost-1:8081/*.txt', 
             'gpfdist://etlhost-2:8082/*.txt')
   FORMAT 'TEXT' ( DELIMITER '|' NULL ' ')
   LOG ERRORS SEGMENT REJECT LIMIT 5;

To create the readable ext_expenses table from CSV-formatted text files:

=# CREATE EXTERNAL TABLE ext_expenses ( name text, 
   date date,  amount float4, category text, desc1 text ) 
   LOCATION ('gpfdist://etlhost-1:8081/*.txt', 
             'gpfdist://etlhost-2:8082/*.txt')
   FORMAT 'CSV' ( DELIMITER ',' )
   LOG ERRORS SEGMENT REJECT LIMIT 5;