For Parquet you might be able to: SET PARQUET_FALLBACK_SCHEMA_RESOLUTION=NAME; (Please look in the documentation for details)
For TEXT, the columns in the table schema are matched to columns in the CSV by else. No other way to do it because TEXT files do not contain schema typically. Your CSV might contain a header with the column names, but Impala cannot use that for resolution. On Wed, Oct 11, 2017 at 11:52 PM, sky <[email protected]> wrote: > Textfile and parquet,the two format both cause data confusion. > > > > . > > > > > At 2017-10-12 14:41:02, "Alexander Behm" <[email protected]> wrote: > >What's the file format? > > > >On Wed, Oct 11, 2017 at 11:30 PM, sky <[email protected]> wrote: > > > >> Hi all, > >> After using the 'alter table ... drop columns ...' to delete the > middle > >> column, then the select query will appear data confusion, how to solve > it ? >
