Confirmed users
574
edits
DXR Storages: Difference between revisions
Jump to navigation
Jump to search
→PostgreSQL: Looked into getting Postgres to do what we want.
(→PostgreSQL: Looked into getting Postgres to do what we want.) |
|||
| Line 41: | Line 41: | ||
== PostgreSQL == | == PostgreSQL == | ||
Postgres is the other contender | Postgres is the other contender. | ||
Pros: | Pros: | ||
* An easy port from our existing relational-based storage | * An easy port from our existing relational-based storage | ||
* | * As of 9.3, out-of-the-box [https://wiki.postgresql.org/images/6/6c/Index_support_for_regular_expression_search.pdf crazy-smart extraction of trigrams from regexes] (see page 49) and accelerated regex searching, comparable to Google Code Search. We'd have to do the extraction in Python (which could be fun) to use ES. Keep in mind that [http://www.postgresql.org/message-id/19399.1299858408@sss.pgh.pa.us pg_trgm's idea of trigrams is, like trilite's, 3 bytes], not 3 chars. Also, they hard-code a limitation to word-like chars, so we'd have to hack the source (undefining the KEEPONLYALNUM macro) to get around that. Even so, [http://code.ohloh.net/file?fid=-_LhALACd35jJVPY3WBWQBo4q-0&cid=eDOmLT58hyw&s=&fp=305491&mp&projSelected=true#L49 they still skip spaces]. We would pretty much have to fork pg_trgm. | ||
== Trilite Unique Advantages == | == Trilite Unique Advantages == | ||