Looks like there is a test failing because of a missing sort-order (and globs can be read in any order - it is file-system dependent). Could you add a rowsort or change the glob to a list of files?

Thanks!

Hi,

Saw the updates here and that #6074 was closed. I checked out head, rebuilt, ran unittest.exe with all tests passing. I then tried my example from that issue and still seeing issue, though with better error message :-)

duckdb csv.duckdb
v0.6.2-dev2340 c0907528a3
Enter ".help" for usage hints.
D .read test.sql
*****************************************************
test sql to read in csv files

SQLite v0.6.2-dev2340 c0907528a3
msvc-1929
Run Time (s): real 0.040 user 0.031250 sys 0.000000
Error: near line 12: Invalid Input Error: Could not convert string 'row_id' to INT64 at line 60338 in column 0. Parser options:
  file=0.csv
  delimiter=','
  quote='"' (default)
  escape='"' (default)
  header=1
  sample_size=18446744073709549568
  ignore_errors=0
  all_varchar=0

Sorry if I jumped the gun on giving this a try...

Hi,

@pdet I spent a little more time looking at the fix and saw my test code became a test case test/sql/copy/csv/parallel/test_multiple_files.test :-)

I took a look at the file and the two tests there both use read_csv_auto() and that does seem to work both parallel and not. However, the issue I was seeing was with read_csv() which is what caused the error in my previous comment. I updated the test file adding:

select * from read_csv('test/sql/copy/csv/data/auto/glob/[0-9].csv', AUTO_DETECT=true)

and that worked. I then added something closer to what failed in previous comment:

`select * from read_csv('test/sql/copy/csv/data/auto/glob/[0-9].csv', sample_size=-1, header=True, columns={'row_id':'BIGINT','integer':'INTEGER','float':'DOUBLE', 'text':'VARCHAR'})'

and that still fails. If I SET experimental_parallel_csv=false then that works just fine. So given it works with parallel turned off I'm assuming the command is valid.

I've attached a modified version of test_multiple_files.test with .txt extension added so that I could attach it. It adds a couple of tests as described above and also adds a foreach loop to cycle experimental_parallel_csv first to false where everything works, and then to true where one test fails.

test_multiple_files.test.txt

Hopefully that is helpful...

Could you perhaps open a new issue or re-open your existing issue? It is easy to lose track of these reports otherwise.

Yeah I get that for sure....problem is I'm not sure how to re-open #6074 again as I'm not seeing any options on how to do that. I was hoping that mentioning it (@pdet mentioned it at the beginning of this pull request too) would provide enough breadcrumbs... Anyway, I'll look a little harder and if I don't have any luck will submit something new.

I have reopened it