Closed mrigger closed 2 years ago
I can cut a release by next week to support SQL.
There are so many dialects. Is there a universal syntax for SQLs?
That would be awesome! Unfortunately, there is no universal SQL grammar. Perhaps supporting the SQLite grammar might be a good starting point?
Sounds good. I will let you know when it is done.
Curious, do you want to use perses for reducing bugs, or as a baseline for evaluation?
For now, I'm mainly interested in using Perses to reduce SQL test cases for a study. Typically, I use C-Reduce + manual reduction, which works reasonably well. There are some patterns that C-Reduce fails to reduce. In the past, I reduced them manually, but for the current use case (a larger evaluation), this is not feasible.
I also have had some thoughts on a new SQL-specific reducer, so having Perses as another baseline would be an additional benefit.
I cut a release. As there are many dialects, you have to name the input file with the extension .sqlite
.
You might also want to enable --enable-token-slicer true --enable-tree-slicer true
to further reduce the input file.
Any feedback is appreciated.
Great, thanks a lot! I'll try and let you know.
It would be great to add SQL support to Perses (e.g., I noticed that several SQL dialects are available at https://github.com/antlr/grammars-v4/tree/master/sql). What would be the estimated effort for adding a dialect?