Issues with Data Preprocessing

[nhatvu205]

I tried to download the data from Google Drive and put it into correct dir for the generate_training_data.py to work properly. However, when I ran the code with this sentence:
python generate_training_data.py --output_dir=data/METR-LA --traffic_df_filename=data/metr-la.h5
It threw an error like this:
Traceback (most recent call last): File "E:\UIT\Nam-3\Ki-2\big-data-analytics\repo\generate_training_data.py", line 101, in <module> generate_train_val_test(args) File "E:\UIT\Nam-3\Ki-2\big-data-analytics\repo\generate_training_data.py", line 48, in generate_train_val_test df = pd.read_hdf(args.traffic_df_filename) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "E:\UIT\Nam-3\Ki-2\big-data-analytics\repo\.venv\Lib\site-packages\pandas\io\pytables.py", line 463, in read_hdf return store.select( ^^^^^^^^^^^^^ File "E:\UIT\Nam-3\Ki-2\big-data-analytics\repo\.venv\Lib\site-packages\pandas\io\pytables.py", line 915, in select s = self._create_storer(group) ^^^^^^^^^^^^^^^^^^^^^^^^^^ File "E:\UIT\Nam-3\Ki-2\big-data-analytics\repo\.venv\Lib\site-packages\pandas\io\pytables.py", line 1866, in _create_storer if "table" not in pt: ^^^^^^^^^^^^^^^^^ TypeError: a bytes-like object is required, not 'str'
Has anyone ever met this problem? Hope this can reach to any can help me. Many thanks!

[gcalistro]

Idk if this is the answer but, when I was creating my own hdf5 files as well as using the original data an issue that came up was the attributes of the data types as well as the metadata in the original file itself. This approach may be more convoluted but for visual and just interactive purposes I used ViTables to open my hdf5 files and make any needed modifications. Its an old platform but it works and felt very simple and user friendly.

[nhatvu205]

Yes, I found the issue. It's about the data itself and version conflict of tables package. Thank you for the solution mate!