What happens if I re-upload a dataset with the same name via the UI?

Re-uploading a file via the UI with the same name as an existing dataset will perform an upsert operation by default. This means that the new data will be merged with the old data under the following conditions:

• If you upload a dataset named “source2.csv” and then upload another CSV file with the same name and schema but different data, the ID of the dataset does not change, and the data from the second upload gets appended to the data from the first upload.
• If columns are missing in a subsequent upload, they are filled with nulls.
• If new columns are present in a subsequent upload, they are ignored.
• If the Primary Keys are the same, "upsert" happens (only one record will exist for each unique primary key value). Please see Understanding Primary Keys for relevant details.
• If the uploaded file is missing the ID column and does not select an ID column, it will generate sequential IDs for the new rows.
• If the uploaded file is missing the ID column and selects some other column as the ID column, the values from the specified column are copied into the ID column