I built a loop to read in KNIME tables (9 tables) which are not extremely big (300.000x12) and it takes ages (3 hours for 1.7 tables) ! If I use 9 separate table readers and concatenate the whole stuff I'm done in a matter of minutes. Why is looping soooooo much slower?
I may not be of a much help, but looping generaly really feels slower. But I am having trouble understanding the whole picture - could you post a picture of both workflows for reference?
sure here it is. It is a very big difference between the two methods. If you want I can put the data on an iDisk for you to upload.
I am unable to reproduce your problem. I have created 12 big tables (300K rows with 21 columns) filled with random values. Saved to disk and then read with loop and concatenate flows like you did. Both instances ran fast and finished in a matter of 1 minute.
If you would like you can post your data (+workflow) and I can try running it and we will see how fast it runs.
Meanwhile, can you state your computer specifications, OS and number of threads KNIME can use (Preferences->KNIME).