Make pernia unfollow you with this one simple trick.
Timeline
Post
Remote status
Context
6
@phnt be correct at him.
@f0x Correct, that's how it happened probably.
@phnt @f0x dude, you are actually just strawmanning me.
i know that the fact its a jsonb structure has nothing to do with the size. the *size* of the row is what makes the index huge.
this is what i'm talking about. you're actually just bullshitting me. and you do this with every single question (that i ask you in good faith) at you.
i know that the fact its a jsonb structure has nothing to do with the size. the *size* of the row is what makes the index huge.
this is what i'm talking about. you're actually just bullshitting me. and you do this with every single question (that i ask you in good faith) at you.
@pernia @f0x
>the *size* of the row is what makes the index huge
This in fact for the third time is not the case. The size of the index is influenced by the size of the indexed part and how many rows the index has. You can have a 5MB row to index, index a 2048 byte part of it and the entry in the index would be the same size if the row was 5KB. Why would an index size be influenced by something it never indexes?
Don't believe me, I've attached a clanker answer as well.
image.png
image.png
>the *size* of the row is what makes the index huge
This in fact for the third time is not the case. The size of the index is influenced by the size of the indexed part and how many rows the index has. You can have a 5MB row to index, index a 2048 byte part of it and the entry in the index would be the same size if the row was 5KB. Why would an index size be influenced by something it never indexes?
Don't believe me, I've attached a clanker answer as well.
image.png
image.png
Replies
4
@phnt @f0x then i misread, i thought the index "activities_actor__data______type___text_id_DESC_NULLS_LAST_inde" btree (actor, (data ->> 'type'::text), id DESC NULLS LAST)
indexed the entire data field, not just data->>'type'
which again, instead of strawmanning and calling me stupid, you could've given me the response you just did now. where was this before?
and it begs the question: how come mitra doesn't have indexes half the size of the rest of the DB, and it runs just as fast?
@pernia @f0x I never called you stupid in that thread related to this and I even gave you a similar answer about the indexes except the clanker answer. But instead you lost your mind for no reason at that reply. And Pete even gave the correct answer why some of the indexes are so big compared to Mitra right under my reply you lost your mind on.
I did call you a mongoloid 14 days prior though, when you complained that a DELETE didn't reclaim space without a VACUUM FULL.
I did call you a mongoloid 14 days prior though, when you complained that a DELETE didn't reclaim space without a VACUUM FULL.
@phnt @f0x yes, for the vacuum thing it was warranted (lol).
perhaps picrel, when you completely blew me off, in the cute and smug little way that is common in microblogging platforms, i took it as a jab.
and then when i had to make a big ass post detailing that yes, mitra's post table *is* normalized, and that yes, i *did* look at it (you fucking dick), it was a little frustrating.
and it didn't help that *you* were then wrong and pete had to correct you (lmao)
it fucking sucks cuz i don't pretend i'm not retarded. i'm pretty transparent. you shit on me for making a guess and you don't even have the answer.
and that's why i unfollowed you.
image.png
perhaps picrel, when you completely blew me off, in the cute and smug little way that is common in microblogging platforms, i took it as a jab.
and then when i had to make a big ass post detailing that yes, mitra's post table *is* normalized, and that yes, i *did* look at it (you fucking dick), it was a little frustrating.
and it didn't help that *you* were then wrong and pete had to correct you (lmao)
it fucking sucks cuz i don't pretend i'm not retarded. i'm pretty transparent. you shit on me for making a guess and you don't even have the answer.
and that's why i unfollowed you.
image.png
