We have a schema where a table has 90 columns.
We have read request patterns where say
api 1 reads column 1-20
api 2 reads column 20-40
api 3 reads column 40-60
api 4 reads column 60-90
And say, we have a write access pattern on the same corresponding group of columns.
Do the write pattern (overlapping/ non overlapping group of columns) affect read performance?
For better performance, should we split the table into 4 sub tables, where each table is used to serve respective read api?