James Prince

I'm not sure what URL you are injecting or how, but to make this work you need to inject **both** the WebSocket URL and the Public API URL. I think...

So I did a bit more investigation. Because of the low number of orders that are submitted to the sandboxes, the set up code as is didn't work if you...

I came across this and tried to recreate it, I've got a snippet which exhibits a similar time profile, the following runs in 12s on my M1 Macbook Pro: ```...

Using the above example, replacing ``` df = df.replace({True: 1, False: 0}) ``` with ``` bool_df = df.select_dtypes("bool") df[bool_df.columns] = bool_df.replace({True: 1, False: 0}) ``` Brings the time down to...

After more digging I think I'm understanding what's going on a little more. The partition filter is resolved higher up the callstack (i.e. not on the worker), and at the...

Seems like a trivial fix to change line 535 mentioned in the comment above to read: ``` if set(predicate_fields).issubset(dataset.partitions.partition_names): ``` It works for my example but this feels far too...

I've opened this PR as a starting point: https://github.com/uber/petastorm/pull/488 I've not been able to get the relevant tests to run locally yet, so maybe Travis will give me some feedback...

I get the following response when I run your code: ``` 1 Failed download: - TSLA: 1m data not available for startTime=1601506800 and endTime=1604966400. Only 7 days worth of 1m...