Let's assume a stream is sunk until block 100000, as stated in issue #2, the latest file is 80000-90000.jsonl.
When resuming the sinking process to get the newest data (giving a new stop range e.g 120000), the latest file from the previous extraction (80000-90000.jsonl) is overwritten by an empty file.
Note: this behavior only appears when the output file path is a local path. (Sinking data to S3 doesn't overwrite the latest file)
Let's assume a stream is sunk until block 100000, as stated in issue #2, the latest file is 80000-90000.jsonl.
When resuming the sinking process to get the newest data (giving a new stop range e.g 120000), the latest file from the previous extraction (80000-90000.jsonl) is overwritten by an empty file.
Note: this behavior only appears when the output file path is a local path. (Sinking data to S3 doesn't overwrite the latest file)