recently turned 21🥂
from being a kid playing csgo at 1am to an adult working on my own product🤌🏻
cannot believe how much ive changed within a yr
a long way ahead, yet to conquer!
thank you all for your wonderful wishes ♥️
Always felt the urge but I guess I need to be precise and accurate enough The only thing I dont like is the limit I can post per conversation...Okay Internals on the go
Postgres is that thing that u cant stop after writing 2 lines Will write on pg_trgm
@SumitM_X In psql the idea main is that we create a new row with updated data stored in some page and mark the existing row as old version so when we run vacuum the old row or dead tuple get removed
ps:-tuple is record that stores page+page offset
kind of part of mvcc
@SumitM_X Never was inappropriate here for sure Still surely can use uuidv7 which is based on time based ordering which makes your b+ tree index tree more uniform spread Surely may not use uuidv4 but still it depends on the use case
@ashoKumar89 uuidv7-creates id based on timestamps and order uuidv4-just generates random sequences as id
Now if u have a to generate id for your table in the db where this id is your primary key or have a index on fields,always go with uuidv7 as uuidv4 is random so your b+tree get skewed
@NitinthisSide_ So I had a doubt so let's say u have created a index on name field in a table and when u search for parameter like "%ash%" using a where query for all the records int the table,now would the executor would use the index or would i it prefer to do sequential scan
I will start writing on internals and working of Postgres Lets see how it goes...
Thinking to start writing on why mvcc is crucial in a certain aspect and which critical concurrency problems it solved wrt to reads and writes in a transaction and also what problems it introduced