cgivre edited a comment on issue #1978: DRILL-7578: HDF5 Metadata Queries Fail with Large Files URL: https://github.com/apache/drill/pull/1978#issuecomment-585272602 @vvysotskyi Thanks for your responses and feedback. Maybe it would help if you saw how this worked: ``` apache drill> select * 2..semicolon> from dfs.test.`dset.h5`; +-------+-----------+-----------+--------------------------------------------------------------------------+ | path | data_type | file_name | int_data | +-------+-----------+-----------+--------------------------------------------------------------------------+ | /dset | DATASET | dset.h5 | [[1,2,3,4,5,6],[7,8,9,10,11,12],[13,14,15,16,17,18],[19,20,21,22,23,24]] | +-------+-----------+-----------+--------------------------------------------------------------------------+ ``` The column that is in question is the `int_data` column. What's happening is if that column is greater than 16MB, Drill runs out of memory. If a user is actually interested in analyzing that data, they should be using a data query which looks like this: ``` apache drill> SELECT * FROM table(dfs.test.`dset.h5` (type => 'hdf5', defaultPath => '/dset')); +-----------+-----------+-----------+-----------+-----------+-----------+ | int_col_0 | int_col_1 | int_col_2 | int_col_3 | int_col_4 | int_col_5 | +-----------+-----------+-----------+-----------+-----------+-----------+ | 1 | 2 | 3 | 4 | 5 | 6 | | 7 | 8 | 9 | 10 | 11 | 12 | | 13 | 14 | 15 | 16 | 17 | 18 | | 19 | 20 | 21 | 22 | 23 | 24 | +-----------+-----------+-----------+-----------+-----------+-----------+ 4 rows selected (0.172 seconds) ``` My point here is that there's really no practical purpose for using this data in a query other than perhaps to see if it is populated at all. A user certainly wouldn't want to aggregate it. All that is happening in this PR is changing the default behavior in the metadata view. I set the default to `false` so that a user would explicitly have to enable this capability. If you're not liking this option, what would you suggest as an alternative. I don't think the current situation is ideal where Drill runs out of memory and crashes...
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
