Use temporary file in load_table_from_dataframe#7545
Use temporary file in load_table_from_dataframe#7545tswast merged 3 commits intogoogleapis:masterfrom
Conversation
This fixes a bug where `load_table_from_dataframe` could not be used with the `fastparquet` library. It should also use less memory when uploading large dataframes.
|
It might be a good idea to have tests that explicitly set the parquet engine to |
|
It looks like there's an error when building snappy |
|
I think I need to add an OS-level package (libsnappy-dev according to StackOverflow) to the trampoline image used for the google-cloud-python BigQuery tests. I see from the Kokoro configs that the image is |
|
Trying a rebuild now that I we have snappy in the Kokoro Dockerfile (googleapis/testing-infra-docker#7). |
|
|
|
The BigQuery tests do pass now. |
This fixes a bug where
load_table_from_dataframecould not be usedwith the
fastparquetlibrary. It should also use less memory whenuploading large dataframes.
Fixes #7543