Unlock a world of possibilities! Login now and discover the exclusive benefits awaiting you.
Hi,
I would like to know if using precedent load is good for performance (executing time)
LOAD *, ..............;
LOAD .....
FROM ...........
I think precedent load is useful to do transformation in the raw data, but how about performance?
There are different loads , how ot works?, how flow the data?
Does it pass from one load to the next when the whole dataset is obtained or it flows row by row (when a row is obtained it passed to the next load although there isn't the whole dataset obtained?
Thanks,
I think this couldn't be answered in general. There are too many possibilities each with benefits and disadvantages. At first a (preceeding) load is really rarely isolated to himself - quite often the load is embedded in a chain of load-statements which are frequently part from a bigger datamodel parted in multiple generators, datamodels and applications. Therefore it's more important to have the overall performance in view and that the script is clear and readable instead to concentrate on a few percent of optimizing by certain loadings (unless you have a somewhere a critical situation where each percent point counts).
Data Modelling: Clarity vs. Speed
In respect to the above mentioned clarity the preceeding load is good tool to have short and logical statements and that it could have longer run-times in some preceeding statements which might to include some where- or group by or order by parts must not be a problem.
Conclusion: I think you will need to test the speed of loadings nearly always to be sure which kind is faster.
Influencing factors for performance of preceding load
- Marcus
Juan how are you?
I pass this interesting blog by hic which explains step by step everything that has to do with the previous load.
Preceding Load
I mention the final phrase Hic done to end this blog
"The Preceding Load Disadvantages have not. Use it. You'll love it.