Re: Is it possible to do incremental load without qvds
I don't really understand your points above.
The purpose of incremental loading is to avoid loading the same data from the source (database or files) over and over, but to store that data in a local file. This is useful if the source file or table is extremely large, or there are access restrictions (eg no expensive reads from a production database during working hours).
You can use any data storage format. But QVDs have some important benefits:
They are compact - although not compressed, QVD's store the data in a very concise format, so the files are smaller (compared to CSVs).
They can be written quickly.
They can be read extremely quickly - eaily 10x faster than reading from a CSV.
They maintain the data types, including dual values. So there are less issues with dates and other complex formats.
Hope that helps clarify the matter.
Logic will get you from a to b. Imagination will take you everywhere. - A Einstein