Python Adding New JSONL Data to an Existing Vertex AI Dataset Video Classification.

[GirishKumarSR]

Hello, I am currently encountering an issue. I have created a dataset and am importing JSONL data from a GCS bucket. After importing, every data point is labeled appropriately, which is fine. However, the problem arises when I insert new JSONL data into the same dataset that was created previously; it shows up as unlabeled in the console. The format of the previous JSONL data and the new JSONL data schema are the same. Additionally, if I import JSONL data from the GCS into an empty dataset, it is labeled correctly. But when I insert new data with the same format, it appears as unlabeled.

I am facing this issue while implementing the Python code and have reviewed some Google Docs, such as:

• https://cloud.google.com/vertex-ai/docs/video-data/classification/create-dataset#aiplatform_create_d...
• https://cloud.google.com/vertex-ai/docs/samples/aiplatform-import-data-video-classification-sample#a..., etc.

In these links, the focus is on inserting JSONL data into an empty dataset, but I need to insert data into an already created and populated Vertex AI dataset.

If anybody has a solution to the above issue, please share the code or any relevant links. Thank you.

[nceniza]

There might be inconsistencies within the values of the json file that you are appending, It might be helpful to post a sample of the added values and the original jsonl file for comaprison.  

[GirishKumarSR]

This is the sample JSONL file I am going to insert into an existing dataset. It contains only two JSON inputs. When I inserted this JSONL data into the existing dataset, which already contains data through Python code, it is being recognized as unlabeled.