SQLite Forum

end gaming to have math extension in pythno3.10 Sqlite
Login

end gaming to have math extension in pythno3.10 Sqlite

(1) By stonebig on 2020-05-09 10:03:44 [link] [source]

Hi all,

I would like to know if there is now a chance a few Math functions (like in "extension-functions") to land in the SQLite official amalgation ?

After Common Table Expressions , Window functions, json functions, it would look like the next logical big thing to bring to SQLite.

Apparently the first step of a public-domain licence in "extension-functions" has happened two years ago

https://gitlab.com/liamh/extension-functions/-/commit/16c2a17c7d9675ecbc717cad1883df21011a71e4

(2.1) By Ryan Smith (cuz) on 2020-05-09 11:25:06 edited from 2.0 in reply to 1 [source]

The big problem with math functions is that they are a niche. (/me avoids glares from mathematicians).

They are very helpful, I compile-in and even add some of my own continuously in SQLite, but not every DB needs them, especially not something which is more a document store or application file format (a big use-case for SQLie).

Add to that the fact they add weight, again not everyone's problem, but certainly some people (embedded etc.) hate bloat.

This is not to say more math functions /won't/ make it into sqlite core, but I mean, it's just one more include in your tool-chain currently, how much easier can it get?

Perhaps you are not a self-roller and lobbying for (and we've seen this request before - and it gets my vote too) one more set of pre-compiled binaries compiled with EVERYTHING, comes standard with all included from RTrees to FTS-Latest to Math functions - to added pragmas, etc. etc. Nice and fat, great for learning/experimenting/using in any situation where a non-dedicated solution is needed.

It'd be nice - but probably a lot of work for a niche case.

[EDIT}: Since posting I've also realized that my math includes are not even coming from the sqlite site, and more importantly, the SQLite team is not maintaining the math function library (which renders some of what I said above moot), and to get it into amalgamation would mean taking it out of the hands of the original devs and more pressingly, a whole new test harness and opening up to new CVE's and fuzzer fails, requiring time and effort to also maintain.

Perhaps they will do it, but it seems like a very high effort-to-pleasure ratio.

(3) By stonebig on 2020-05-09 18:32:50 in reply to 2.1 [link] [source]

Since "extension-functions" was created (in 2010 ?):

  • SQLite integrated the string handling parts,
  • SQL:2016 apparently added to the standard the trigonometric functions.

This would make SQLite show better in comparison sheets like in https://modern-sql.com/blog/2017-06/whats-new-in-sql-2016 , and make it more relevant as a neutral/public domain support for teaching SQL.