Skip to content
This repository was archived by the owner on Jan 28, 2021. It is now read-only.
This repository was archived by the owner on Jan 28, 2021. It is now read-only.

Indexes on language func does not work with aliased cols #639

Closed
@dpordomingo

Description

@dpordomingo

discovered at https://github.com/src-d/empathy-sessions/issues/37#issuecomment-470041061

time srcd sql "SELECT file_path, language(file_path, blob_content)
FROM files
WHERE language(file_path, blob_content)='JSX';"
+------------------------------------+--------------------------------+
|             FILE PATH              |   LANGUAGE(FILES FILE PATH,    |
|                                    |      FILES BLOB CONTENT)       |
+------------------------------------+--------------------------------+
| src/js/components/posts.jsx        | JSX                            |
| src/js/components/repositories.jsx | JSX                            |
+------------------------------------+--------------------------------+
real	0m12.926s

Then I created an index, following your example to create an index on language expresion

$ time srcd sql "CREATE INDEX files_lang_idx ON files USING pilosa (language(file_path, blob_content));
+
+

real	0m0.236s

And queried again, and the time processing was reduded from ~15s to ~.3s what is nice.

But then I queried using an aliased column... and ~15s again.

$ time srcd sql "SELECT file_path, language(file_path, blob_content) as lang 
FROM files
WHERE lang='JSX';"

...

real	0m12.147s

Metadata

Metadata

Assignees

Labels

bugSomething isn't workingempathy-sessionsIssue filed as as part of empathy sessions

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions