Bulat Yapparov

@m-rtijn I am not sure if I am doing something wrong, but I get bias of 16 d/s on the X gyro data. that makes it pretty much unusable without...

@risadams This should be reopened unless it is a duplicate. Request seems to be pretty clear and relevant, unless there is an alternative way to achieve this.

You have double layer of timepartitoning lists. Can you double check the example from the previous issue

@SeagleLiu, Just note that you can already create an empty table with partition using `bigrquery::bq_table_create()`: ```r clustering

@philbrierley, do you posing repex with data to make it easier to investigate this. Also, an you clarify >The same data uploaded by other libraries has no such issue. Any...

ok. I can reproduce this, with both `data.frame` and `data.table`. this rounding is done when data is transformed to json, see ratio values: ```r cars$ratio

I think it is easier to implement something like this externally from the package, as a wrapper that combines functionality of `bigrquery` and `googleCloudStorageR` packages. See example of a wrapper...

@luiu1234 can you try version of bigrquery from my PR to confirm it is doing what you were asking for.

@hadley I have used this documentation to implement it: https://rstudio.github.io/rstudio-extensions/connections-contract.html BigQuery objects need to be linked to this contract, which kind of makes sense. Let me know if you know...

@jennybc I don't think I see what you see. When I look at changes in GitHub, all seems relevant to this PR.