I need to run asof join on few years data of trade and quote tables which are partitioned.
When I read through - Above url states that -
"If further where constraints are used, the columns will be copied instead of mapped into memory, slowing down the join." How can I use asof join over partitioned database with date and other constraints without impacting performance or memory.
Eg: aj[`sym`time;select from trade where date>2019.01.01, app=`abc; select from quote where date>2019.01.01] 1 Answer
You should do date= and run for each/peach date rather than doing date>, i.e.
raze{aj[`sym`time;select from trade where date=x, app=`abc;select from quote where date=x]}peach 2019.01.01 2019.01.02 2019.01.03 It's usually acceptable to have an additional filter on trade but probably only if that column app has a parted attribute. You can't have any filters on quote or performance tanks.
Note that with this approach you can't join prevailing data from previous day onto the next day but in most cases you wouldn't want/need to do that anyway
2ncG1vNJzZmirpJawrLvVnqmfpJ%2Bse6S7zGiorp2jqbawutJobHJsYWmBc3%2BOoaawZaSkeqK8z6WwZpmjpLNuts6ipWannmK9or7Toquip56asW7AwJujnqtdnrtut8Ob
