I have a requirement to store some huge collection of data just until I consume it. For example, I download a file(a csv probably, which can be some hundreds of MB) and store its contents in a db. After the whole file's written to db, another process read this collection and delete it.
I prefer using MongoDB for this. However I was told VoldDb is better for this kind of purposes.
I would like to know why exactly VoldDB is and how its better(if it is) than mongoDB for temporary storage.
This sounds like a batch-driven process with static data. While you could do it with VoltDB, it would also be capable of ingesting the data in real-time from the source or sources rather than after it has been accumulated into a CSV file, and rather than simply buffering the records, you could do some processing, validation, or transformation of the data as it is ingested per-record inside of a transaction, and then you could use export to stream the output to various target systems.
That being said, there is a CSV Loader you could use to bulk load the csv files, and then the other program can retrieve the data using one of the various client libraries.