Importing large CSVs and consolidating in one matrix.

Hello I need to consolidate around 800 CSVs(total size ~ 6GB) into 1 matrix. I need to process(mainly filtering) the data and plot the graphs. What should be the approach of the task so that the processing time in loading the CSVs and plotting takes minimum time.

TIA

Answer 1

[Titus Edelhofer]

Hi Pranjal,

if you have R2014b available, then datastore + mapreduce would be the choice of tools.

Titus

[Pranjal]

Hi Titus,

Thanks for the answer,but I am working on R2011b. Can you suggest any approach using that version of matlab.

Pranjal

[Titus Edelhofer]

In this case use textscan for reading the data. Each file is probably not too large (assuming the 800 csv are somewhat similar in size), i.e., about 7-10 MByte. Therefore reading each file in one sweep should be no problem.

You might use the Parallel Computing Toolbox (parfor) to read the files in parallel. Either process the data right away or save to .mat so that you can read them later faster in a loop.

Titus