I’m running a query that pulls an entire board which contains about 400 items and 80 columns. i’m using a pretty blunt query:
query2 = ‘{boards(ids:###) { name items { name group{title} created_at column_values{ title text } } } }’
It takes about 20 seconds for this to run, which is surprising given that this is just tabular text data. Is this normal latency? I would expect a much faster response on this (maybe 1s). I’ve seen a few (unaddressed) threads about performance of the API - is this an ongoing issue?
I have found that the most significant factor in API response reading items is in returning connect boards and mirrored columns.
80 columns is a lot. If there are any of these column type that you do not need, I would expect that removing them would significantly improve your response time.
It’s hard for me to see how 80 columns of text data (maybe 100B per row) is ‘a lot’ in terms of data transfer. there are api’s out there serving up enormous amounts of data in near real time.
My comment he about the 80 columns was from a human perspective not a data transfer perspective. The reason I made it was because I was imagining having to include the list of columns by name in your query.
The point of my post was that mirror columns (and possibly connect boards) are significant factor in the response time of the API. That is all.
Thanks Jim.
Since I’m calling column_values in one statement i’m not having to list out all 80 columns. You’re right, 80 is a lot of info to sort through, but I find its easier for me to pull it all into python and work with it there than to identify the columns i need from the board initially.
I will check on those mirrored columns as you say, maybe i can eliminate a few.
I just wanted to comment here as you are not the only one that has issues with the seemingly super long query times. Unless you are getting a single item, it is really, really, really slow in comparison to other companies.
My experience with the API has been significantely slower than others I’ve used in the past. Hopefully they are working on fixes here as it is extremely slow for companies that actually use this for any type of CRM purposes.