Repository navigation
Reading table with URL is doing many requests - #2001
Conversation
|
Huh, that's not good! |
|
I haven't looked at how it works, but I ran the commands again (this time in master, not 0.3) and it seems like there is maybe a default timeout that is too short. Using master, after trying ~10 times it actually gave a timeout exception. |
|
Why would the log message read "[Done]" if the request was actually timing out, or is that an oversight too? :) |
|
I did it a slightly different way and did not see such behavior. FYI. |
|
I double checked and it seems this is still an issue. My guess (pun not intended) is that it's due to the guessing process in io.ascii, except it's downloading the file every time instead of using a cached copy. @pllim's version works because by using |
|
Code attached. I haven't read the @embray - you were right about this being related to guessing. cc: @hamogu |
|
I don't really understand the fix, but it works for me so 👍 |
|
This is a good way to do it (I think I can say that after spending significant time on So 👍 to merging this. Just for the record: Not all CDS tables have a separate |
Reading table with URL is doing many requests
Reading table with URL is doing many requests
If I read a web-based CSV file using
requestsit works straightaway and quickly:But if I do what I think is the equivalent directly with
Table.readit's hitting the site many times and takes probably 10 times as long: