If I ran a simple ‘SELECT * FROM “folder”.“v2.6”’ query on a folder with many csv files, I get the result as expected.
If I ran this query, ‘SELECT dir0, dir1, dir2, COUNT(*) AS Count_Star FROM “folder”.“v2.6” AS “v2.6” GROUP BY dir0, dir1, dir2’, the preview shows the correct groupings with count 1000 on each grouping, as expected. But when I click Run to run the full query, I get this error. Only one data line detected. Please consider changing line delimiter.
For the “Run” is a job created? If so, can you attach the profile for that job?
When you “Preview” or when you “promote” (format) a folder to a physical dataset, Dremio samples that folder’s files; it doesn’t read every file with the specified format. But when you “Run” the query, Dremio will starting reading all of the text files in the partitions you’ve filtered on your query. If one of them doesn’t have the same format, the error may be produced.
Hi Ben. Thanks for the reply. I managed to track down the culprit, thanks to the profile, to a file that had a empty/null row at the bottom.