I always thought that most \-commands in psql were just frontends for queries? For example:
% psql db_name
(1)=# \set ECHO_HIDDEN
(1)=# \d
********* QUERY **********
SELECT n.nspname as "Schema",
c.relname as "Name",
CASE c.relkind WHEN 'r' THEN 'table' WHEN 'v' THEN 'view' WHEN 'm' THEN 'materialized view' WHEN 'i' THEN 'index' WHEN 'S' TH
EN 'sequence' WHEN 't' THEN 'TOAST table' WHEN 'f' THEN 'foreign table' WHEN 'p' THEN 'partitioned table' WHEN 'I' THEN 'partit
ioned index' END as "Type",
pg_catalog.pg_get_userbyid(c.relowner) as "Owner"
FROM pg_catalog.pg_class c
LEFT JOIN pg_catalog.pg_namespace n ON n.oid = c.relnamespace
LEFT JOIN pg_catalog.pg_am am ON am.oid = c.relam
WHERE c.relkind IN ('r','p','v','m','S','f','')
AND n.nspname <> 'pg_catalog'
AND n.nspname !~ '^pg_toast'
AND n.nspname <> 'information_schema'
AND pg_catalog.pg_table_is_visible(c.oid)
ORDER BY 1,2;
**************************
List of relations
Schema │ Name │ Type │ Owner
────────┼────────────────────────────────────────┼──────────┼────────
...
I re-implemented quite a bit of that in a PostgreSQL management tool I built, based just on the ECHO_HIDDEN queries. Never even had to look at any C code.
I'm not sure if I entirely follow what translating C to JS gives you, other than a bit of (fairly simple) parsing of \cmd and flags? I didn't look too carefully at the code, so maybe I'm missing something.
Some commands issue multiple queries and include logic for stitching the results into a nice table. There’s not a one-to-one mapping from commands to queries by any means.
Yeah, that's right; but the "glue code" I've had to write for this has been really minimal. Certainly less effort than translating C to JavaScript with regexps (that on its own is something of an accomplishment). I didn't port everything so maybe I missed some more complex bits.
I mean, if this works well for you then that's grand; just wanted to point out to others that \-commands are typically very "non-magic" and you don't need to go the route of translating C code.
Almost every single post you make mentions clickhouse. You've replied to a few of my posts now with this vaguely on-topic forceful injection of clickhouse.
Of course I'm not saying you can never bring up clickhouse, but at this point it's just becoming spam. High-effort spam, maybe. But still spam.
The fact that this is not a manual reimplementation, but a "transpilation" of the psql source from C to JavaScript, makes this super interesting to me.
From the Postgres master branch (17devel), we take exec_command_d, exec_command_list and exec_command_sf_sv from command.c, and all of describe.c and sql_help.c, from src/bin/psql.
We use plenty of RegExp search-and-replace to turn this C code into valid JS syntax.
We implement some C library functions, such as strlen and strchr, and some Postgres support functions, such as printTable and printQuery, in JavaScript.
When I started my programming journey I thought I could just "translate" Pascal to C++ just with search/replace and minor adjustments. Turns out sometimes it's the best approach :)
OP (and library author) here. This is a fair point, and one I freely admit to in the accompanying blog post. It’s probably not crucial to sync more than once per major Postgres version, but we’re still figuring out how we’ll deal with this.
Thw convertet is nice, I could imagine it being it's own library, with the generator, and a release including the generated code every postures release.
I always thought that most \-commands in psql were just frontends for queries? For example:
I re-implemented quite a bit of that in a PostgreSQL management tool I built, based just on the ECHO_HIDDEN queries. Never even had to look at any C code.
I'm not sure if I entirely follow what translating C to JS gives you, other than a bit of (fairly simple) parsing of \cmd and flags? I didn't look too carefully at the code, so maybe I'm missing something.
If you just need one or two of the commands, just using those queries directly is easy.
If you want to support all of them, including handling all the different Postgres versions, it becomes quite a lot to implement.
Some commands issue multiple queries and include logic for stitching the results into a nice table. There’s not a one-to-one mapping from commands to queries by any means.
Yeah, that's right; but the "glue code" I've had to write for this has been really minimal. Certainly less effort than translating C to JavaScript with regexps (that on its own is something of an accomplishment). I didn't port everything so maybe I missed some more complex bits.
I mean, if this works well for you then that's grand; just wanted to point out to others that \-commands are typically very "non-magic" and you don't need to go the route of translating C code.
1 reply →
Yeah, I don't understand why they'd translate the C code instead of just writing catalog/info schema queries, which seems easier AND more educational.
[flagged]
Almost every single post you make mentions clickhouse. You've replied to a few of my posts now with this vaguely on-topic forceful injection of clickhouse.
Of course I'm not saying you can never bring up clickhouse, but at this point it's just becoming spam. High-effort spam, maybe. But still spam.
6 replies →
The fact that this is not a manual reimplementation, but a "transpilation" of the psql source from C to JavaScript, makes this super interesting to me.
Read more here: https://neon.tech/blog/bringing-psqls-d-to-your-web-browser
When I started my programming journey I thought I could just "translate" Pascal to C++ just with search/replace and minor adjustments. Turns out sometimes it's the best approach :)
OP and library author here. You’ll find the background to this in this blog post: https://neon.tech/blog/bringing-psqls-d-to-your-web-browser
Ok, we've changed the URL to that from https://github.com/neondatabase/psql-describe above. Thanks!
This seems really cool, but also a nightmare to maintain. What are the plans for keeping this in sync with the upstream codebase?
OP (and library author) here. This is a fair point, and one I freely admit to in the accompanying blog post. It’s probably not crucial to sync more than once per major Postgres version, but we’re still figuring out how we’ll deal with this.
uses emscripten to compile the C to WASM, so it’s automated.
Thw convertet is nice, I could imagine it being it's own library, with the generator, and a release including the generated code every postures release.
I'm not sure I understand how to use that.
It’s used in Neon’s web-based console to support \d and friends alongside regular SQL queries.
You might well never need to use it, but if you’re building an online query editor it could be very handy!
Yeah, I don't really get where / how I would use this. The documentation is really lacking on outlining the usage / use-case.
Anyone using Neon in prod? How is it?
I wish \d allowed sorting alphabetically.