00:26:16 GMT <bleb> minus: and do you know where redis-cli gets it's default connection? i.e. if you don't specify something with -s
04:12:32 GMT <HelloWorld1> Hi all, been a while since I logged onto IRC. Thought I'd try my luck here with a query I have. I am aware of Redis Pub/Sub as well as keyspace notifications. I am wondering how I could go about receiving notifications for a particular key? The problem I have is that a springboot based client using jedis which connects to Redis. When it requests a key, if it does not exist. It should subscribe so that it gets notified when it beco
04:13:41 GMT <HelloWorld1> Another approach that crossed my mind is that I could 'poll' for say 500ms and then give abandon if it's still not there
06:26:32 GMT <minus> bleb: well, localhost:6379 is the default ip:port redis listens on (the values are listed in redis-cli --help)
06:30:38 GMT <minus> HelloWorld1: when it's a key expiration notification then you won't be able to retrieve the key contents afterwards, otherwise it should be fine. depending on what you're trying to do you could push things into a redis list and brpop (blocking pop) on it
08:25:09 GMT <HelloWorld1> @minus, I have key expiration covered. It's just trying to figure out an optimal way of solving my problem. Right now my app hits Redis via a rest api (another springboot app) every few ms. But I don't like this approach. Too much 'chatter'. I want to isolate that between the app that talks to Redis directly
08:31:47 GMT <minus> HelloWorld1: your lines are too long and get cut off
08:32:08 GMT <minus> > …app that talks to Redis directly ← last thing that's there
08:32:47 GMT <minus> hm, that line wasn't even that long
11:07:44 GMT <tarkus> How do you think, should I run Redis docker image under "redis" user in production?
12:12:43 GMT <saml> hey, how do I do HDEL of 1 million?
12:12:56 GMT <saml> i have a line separated file. each line lists id
12:13:32 GMT <saml> redis-cli HDEL rootkey $(cat thefile.txt)
12:13:47 GMT <saml> it'll probably hit my shell's argv limit
12:24:42 GMT <minus> saml: DEL the rootkey?
12:25:19 GMT <minus> also, having 1M entries in a hash probably isn't the best thing to do
12:28:28 GMT <saml> minus, actually HDEL
12:28:42 GMT <saml> ah i see what you mean
12:29:28 GMT <saml> is there HCOUNT ? a way to see how many entries in a hash
12:30:28 GMT <saml> hlen  gives me 300 million entries
12:30:37 GMT <saml> i just want to delete 1 million only
12:31:13 GMT <saml> only 0.3%
12:34:08 GMT <minus> use a client and pipelining then
12:34:22 GMT <minus> 300M entries in a hash sounds very wrong
12:34:27 GMT <minus> that must be rather slow
12:35:18 GMT <minus> otoh HDEL is O(N) for N entries to delete
12:46:39 GMT <saml> minus, client pipelining means multiple invocation of redis-cli?
12:46:50 GMT <saml> redis-cli -x HDEL rootkey < ids.txt
12:47:11 GMT <saml> this does not work as I expected.  tried both line delimited ids.txt or space delimited
12:47:16 GMT <minus> nah, i mean don't use redis-cli
12:47:36 GMT <saml> ah thanks
12:47:54 GMT <minus> like, write a python piece to do the thing
12:48:15 GMT <minus> though i'm not entirely confident on how/if pipelining works in redis-py
12:48:46 GMT <Habbie> redis-py has pipelining
12:49:34 GMT <minus> afaik it has transactions/multi, but not pipelining in a sense that you don't wait for the result in place
12:50:24 GMT <Habbie> https://github.com/andymccurdy/redis-py#pipelines
12:51:23 GMT <minus> yeah, that's what i meant
13:00:41 GMT <saml> purge all the cache
18:19:05 GMT <ced_> Is redis suited for the situation in which I need sorted set as well as one sorted set than contains all the others. Think Ranking by day but also ranking since the beginning
18:21:10 GMT <minus> they're separate data sets, so probably
18:21:54 GMT <minus> might still be a better idea to just drop it in some RDBMS if you need other things as well
18:27:56 GMT <ced_> @minus, thanks I feel like RDBMS might be better suited. The thing that's putting me off is that I'd need to duplicate data in my scenario right ? So I'd have to keep track of the score in the "score_by_day" zset as well as the "score_since_forever" zset
18:28:19 GMT <ced_> That is, for the same data
18:35:44 GMT <minus> depends on how you calculate the score, i guess
18:35:54 GMT <minus> and you can just store references in the zsets
19:28:43 GMT <stickupkid> I wonder if anyone has any advice when implementing redis modules. I want to store a list of strings in a hash (HSET, etc), what's the best way to put those values into the hash field? JSON, msgpack (is that available?) or simple CSV?