01:53:45 GMT <Antaranian> long time no talk
15:03:22 GMT <Kingsy> is there a limit to the amount of characters I can store against a key? Just as a string
15:04:09 GMT <Kingsy> oh... 512MB
15:04:11 GMT <Kingsy> thats MASSIVE
15:19:28 GMT <badboy_> not really^^
15:25:27 GMT <Kingsy> 512MB for a string is pretty big I would think?
15:25:38 GMT <Kingsy> if you think about how much space one char would use
15:27:41 GMT <minus> redis rather stores byte arrays than strings
20:12:43 GMT <Blight_> Hi guys quick question, keyspace_misses increases on what operation GET or SET as well?
20:20:15 GMT <badboy_> GET. as obviously SET can't miss because it doesn't do any lookup
21:09:12 GMT <Blight_> badboy_: internally SET does a lookup no?
21:09:29 GMT <badboy_> no, why should it?
21:11:15 GMT <Blight_> like the XX flag
21:11:31 GMT <Blight_> that update the value only if it exists
21:11:47 GMT <badboy_> ah, well, the more experienced flags would need to do it, yes
21:11:58 GMT <badboy_> I only thought about the simple "SET a b" case
21:12:59 GMT <badboy_> though it seems those are not counted as misses
21:14:11 GMT <badboy_> ah, yeah, the miss counter is only done on purely read operations
21:17:10 GMT <badboy_> (fun how many ways there are to look up a key)
21:18:04 GMT <Blight_> Yeah i'm digging the source as well
21:18:25 GMT <Blight_> there is lookupKeyWrite()
21:18:37 GMT <Blight_> just making sure it's beeing called from SET
21:21:44 GMT <Blight_> So yes it seems that it's beeing updated if I read that correctly https://github.com/antirez/redis/blob/41d16f7a4a6848be304e0b4433bd5683caf27338/src/t_string.c#L80
21:22:09 GMT <Blight_> for NX and XX flag ( SET command )
21:24:17 GMT <badboy_> lookupKeyWrite does not touch the keyspace_miss counter
21:24:36 GMT <badboy_> it is only touched here: https://github.com/antirez/redis/blob/0dbfb1d154b0df28bbdb16a59ae7342d4a3f9281/src/db.c#L122
21:28:35 GMT <Blight_> lookupKeyWrite calls lookupKey
21:28:56 GMT <Blight_> it's just 10 line bellow your link :D https://github.com/antirez/redis/blob/0dbfb1d154b0df28bbdb16a59ae7342d4a3f9281/src/db.c#L139
21:32:54 GMT <badboy_> yes, but as I said: it never touches keyspace_miss
21:32:59 GMT <badboy_> only lookupKeyReadWithFlags does that
21:40:28 GMT <Blight_> lookupKeyRead() also which makes sense
21:41:48 GMT <badboy_> uhm, so?
21:42:01 GMT <badboy_> whatever, it is not called in any way from setCommand
21:42:05 GMT <Blight_> so I'm checking where those functons are called
21:42:12 GMT <Blight_> functions*
21:43:27 GMT <badboy_> let me make it easy for you: the keyspace hit/miss counters are only incremented on read commands
21:48:34 GMT <Blight_> looks like it yes, thank you
21:53:11 GMT <pikajude> hello
21:53:26 GMT <pikajude> i have a bunch of hashes stored in redis, and i need to iterate through them all and delete items whose keys start with a specific string
21:53:33 GMT <pikajude> what's the best way to do that
22:03:52 GMT <badboy_> do you know all keys of those hashes?
22:05:18 GMT <pikajude> i can get them pretty easily
22:05:24 GMT <pikajude> they're all user:*
22:08:44 GMT <minus> so that's a no then
22:09:13 GMT <minus> you can find them using KEYS/SCAN, but that could be bad since it's O(n) on the size of the keyspace
22:11:16 GMT <pikajude> there are about 11,000 keys
22:12:32 GMT <badboy_> if it is just for once do the SCAN variant
22:12:47 GMT <pikajude> yes this is a one shot task to delete some vestigial data