guillesd comments

Results 70 comments of


                                            guillesd

ducklake fails to update table stats on insert into ducklake table on mysql

Hi @alexanderhupfer @ededovic unfortunately we have seen that there are several limitations regarding the DuckDB MySQL connector that are limiting for the DuckDB Ducklake extension. We have placed a warning...

merge_adjacent_files not compacting a table with lots of single row files

Potentially related to this https://github.com/duckdb/duckdb/issues/19552 I'm gonna tag this as needs a reproducer but let's see if this is fixed for 1.4.2

merge_adjacent_files not compacting a table with lots of single row files

Could you provide the reproducer then? I understand what you say and if we decide to open the possibility of merging files with different schemas we will need a nice...

Columns from hive partitioning should be used to skip reading external parquet files

I can't reproduce this. Are you sure those get requests are from this query? `SELECT DISTINCT(crawl) FROM commoncrawl.view;`?

Columns from hive partitioning should be used to skip reading external parquet files

You haven't really specified that the table is partitioned after creation (via `ALTER TABLE`, so when you add the files no partition metadata is saved).

Columns from hive partitioning should be used to skip reading external parquet files

What I meant is that you are not partitioning the table https://ducklake.select/docs/stable/duckdb/advanced_features/partitioning However this doesn't matter because right now we don't seem to be supporting adding partition value information from...

Columns from hive partitioning should be used to skip reading external parquet files

Yes I agree @chaegordon ! Although this issue is a bit missleading I think the point should be --> partition metadata should be populated on `ducklake_add_data_files`

`SELECT COUNT(*)` does unnecessary http requests which slows down large ducklakes

We will at some point expose something, maybe with some existing syntax like `SUMMARIZE`. I think tweaking path of actual `SELECT` queries is not a great option. For now you...

ducklake catalog backup fails on 'Missing Extension Error'

hi @macraesdirtysocks! I think this is a mistake on my side, since the SQL dump backup managed by DuckDB should be between postgres databases rather than a postgres db and...

ducklake catalog backup fails on 'Missing Extension Error'

I added a warning to the page. I think the problem generally with the COPY command is that it doesn't handle edge cases pretty well. That is why we also...