Replace whole data for a key and drop old values

I want to replace whole data for a key every one minute:

Suppose we have a table like this:

Now I want to replace all records with cid=1 with new values which is b and c;
I have two plans:

1. Remove all existing with delete from table where cid=1 and then insert into table values(1,b),(1,c);
2. Insert all new records by insert ignore into table values(1,b),(1,c) and delete old data which is not exist in new values delete from table where cid=1 and uid not in (b,c)

I can use partitioning in mysql on cid which causes fast delete of data but I'm not sure if there is a better method since this happens every minute with thousands of values. I want the best method with minimum time and lowest cost (CPU+DISK+Blueprint). I need something like overwrite a file with new content every minute (We don't need old data anymore).

any suggestion?

If you want to maximise read access and have a full dataset you can load at once, then you might consider a possible third option:

• create an empty clone of the original table

  CREATE TABLE mytable_new LIKE mytable;

• populate your new table with a full dataset. There are different ways you could do this, depending on where your new data is coming from. LOAD DATA is pretty efficient if your data is already on the server, or you could use batches of INSERT INTO mytable_new VALUES (1,'a'),(1,'b'),... statements if your data is coming via an application layer, or, if your new data is already elsewhere within MySQL tables, INSERT INTO mytable_new SELECT ...

• do an atomic rename of your old and new tables

  RENAME TABLE mytable TO mytable_old, mytable_new TO mytable;

• finally, clean up by dropping your old data

  DROP TABLE mytable_old;