00:12 hankache_ joined
SmokeMachine How cool is that? usercontent.irccloud-cdn.com/file/.../image.png 00:16
00:16 hankache left 00:35 Pixi` is now known as Pixi
japhb Pretty cool. :-) 00:53
disbot <simon_sibl> @antononcube on that wasm vm not much, was mostly to see if its possible and how well it run on a slow cpu with 1 thread, but at my work I use it anytime I need to do some scripting, especially since Rakupp got 100% roast coverage, that was my only "pain point" with Raku that its not easy to distribute to people who dont use Raku 01:41
01:56 hulk joined 01:57 kylese left
disbot <simon_sibl> I wish a candidate could start using perl or raku during the interview 01:59
<simon_sibl> instead of google AI searching how to block ping (they dont know its ICMP) on the firewall ;-;
<simon_sibl> especially since they usually have on their cv "firewall", "linux", "security", and security is a big point to me 02:00
<simon_sibl> for me someone who has "security" in their cv should mean that they are curious or at least dig topics more
02:15 hulk left, kylese joined 02:57 hankache joined 03:00 hankache_ left 04:55 Sgeo left 06:06 abraxxa joined, abraxxa left 06:08 hankache left, abraxxa joined 06:26 abraxxa left 06:40 _________ left 06:42 _________ joined 07:40 ShimmerFairy left 07:41 ShimmerFairy joined
Voldenet that RED example is quite impressive, a bit like magic 07:44
but also probably slower than fetching both things from db and doing in-memory join 07:45
because the db has to return data of N^3 size 07:47
so it would allocate this much memory instead of 3N size, which would be in-mem join 07:49
and if db driver is streaming the results, the client has to process the same data for every record most likely 07:51
SmokeMachine you can further filter the results on the query just adding a `.grep` on that `(Model1.^all X Model2.^all).grep: -> ($model1, $model2) { $model1.id2 == $model2.id }` for example 07:59
Voldenet well in that case it makes total sense 08:07
08:07 [Coke] left
Voldenet I've been performance engineering queries for too long I suppose 08:08
ab5tract another reason to avoid in-db transforms is just to keep the pressure off the db server itself. still, this looks super slick! 08:33
08:34 [Coke] joined 08:36 abraxxa joined
Voldenet actually transforms are usually cheap 09:02
event he string manipulations and so on
even the*
just the multiple collection ones are costly
09:11 hankache joined 09:33 abraxxa left 10:19 dustinm` left 10:32 dustinm` joined 11:08 acidsys left 11:09 acidsys joined 12:06 kylese left 12:07 kylese joined 14:12 Sgeo joined
ugexe the transformations themselves can be cheap but depending on where they are they used you might get an inefficient plan 14:35
15:14 corwin left
Voldenet Ah, right, of course, WHERE on transformed row could even go over the whole table 15:32
if it's not pre-computed or indexed 15:33
ugexe it might not even matter if its indexed such as with something like WHERE lower(email) = ? 15:40
(which is just a naive example and how i'd suggest checking for an email) 15:41
s/how/not how/
evalable6 Use of uninitialized value $_ of type Any in string context.
Methods .^name, .raku, .gist, or .say can be used to stringify it to something meaningful.
in block <unit> at /tmp/LihmvB0AC8 line 1
Voldenet depends on your db, mssql can use computed fields with this, postgres can use expression index 15:42
though the expression must be equivalent which can get tricky depending on datatypes 15:43
it's better to add computed or generated column, index it and do query on it directly 15:45
I like how expression/functional indexes can be established with no code change 15:48
15:57 silug left 16:04 silug joined 18:14 stanrifkin joined 18:26 silug left 18:30 silug joined 18:31 sorenson left 19:13 Sgeo left 19:36 Sgeo joined 21:12 guest24 joined 21:13 stanrifkin left 21:14 guest35 joined 21:18 guest24 left 23:33 fdp is now known as Guest7002, Guest7002 left, fdp joined 23:42 hankache_ joined, guest35 left 23:44 hankache left