'COPY ... FROM' incredibly slow before performing query after install


I think there’s some housekeeping missing after installation. I installed via mapd.com/download using Ubuntu x64 > apt instructions. I skipped messing with sample data and went straight to loading my data. It loaded at a rate of about 5900 records per second second. After performing a query to test it out, my next table loaded at over 1 million records per second and I’ve had no issues with slow loading ever since.

Seems like there’s some housekeeping that didn’t get done until I ran the sample query. Just wanted to give a heads up, thanks for an amazing engine!



Glad to have another new user on MapD.

There is little connection between the query execution path and the bulk data load path. Bulk load is all CPU and disk IO.

Nothing springs to mind product wise why there would be any correlation between your observations?

If you stop your MapD server and restart it and load the same file again into a different table do you still see the very slow load times?



I actually can’t get it to be slow again now that I’ve started using it. They were very close together, I did the initial load which was very slow, when it finished I immediately ran a test query on it, then loaded the next table which went extremely fast.

Maybe just coincidence but running the sample query was all I could think of.