CDC of csv and blobs
Opened this issue · 1 comments
I know this is not meant to be performant or adhere to linearizability / serialisation aspects of a db …
also I like the simplicity!!
What if when a row changes we do CDC ( change data capture ) on it?
higher lever tools can react to the CDC event . This brings the unix pipes programming approach. Say a csv file has a list of subscribers to you blog . You can now build a golang binary that responds to build blog articles. Just a simple example.
Also images / videos / blobs are commonly used with data. A csv row can reference an image - it’s quite a typical use case pattern.
But what if the image name changes of is replaced ? CDC of the file system then allows the csv system to react to the change and update the renamed image file keeping everything in sync.
Another use case is the queries for the csv. If you cdc the csv you can see when table or column name changes and update the query in the code. It’s a real refactoring system . Continuous compilation
Interesting concept, thanks for the issue. Is that a common thing, for renaming (on disk) a referenced file (referenced in a row in a DB) and then having that referenced path in the DB change also? Is this a PR you are preparing?