I honestly can’t remember if I’d read somewhere on the website if only RowStore tables are supported for the INSERT IGNORE INTO query, or if both Columnstore & RowStore tables are supported.
I would test myself but currently away from my machine. However, one of the Rowstore tables utilizing the command is quickly taking up RAM resources, so the plan was to migrate that into a Columnstore table at some point.
Just wanted to ensure if I’m still able to use the INSERT IGNORE INTO query for those Columnstore tables.
Currently, anything requiring a unique key only works on rowstores. INSERT IGNORE falls into this category, along with INSERT ON DUP KEY UPDATE. We have plans to address this in our SingleStore roadmap. But you will need to use rowstore or work around it for now.
Using a columnstore hash index and some code in a stored procedure, some people have been able to work around this kind of thing fairly well.
No problem and thanks. I can definitely think of a workaround, using a stored procedure as you’d mentioned that deletes duplicates (while keeping a single copy).