I have a large table of data approximately 200,000 rows by 4000 columns. This is too large for excel and so a collegue pointed me to KNIME
I find it perfect for my needs but I have one problem. I want to search the table for specific entries and then replace each instance with a blank cell. In excel I would do this using the find replace command which has the option to replace all. I have been unable to find anything similar in KNIME
I did try converting every entry to string then setting up a loop so that string replacer would work one column at a time but this is very slow. Is there a quicker way?
Have you tried using the Column Combiner node first using , delimiter and " as Quote character.
Then a Column Filter node to remove the individual columns you just combined.
Then use the String Replacer on this combined column to do the replacements you desire.
Now use the Cell Splitter node with , delimiter and " as Quote Character to get the original columns back.
You will notice the original column names have been lost, if those are important to you, then instead of the Column Filter node earlier, use the Splitter node to split out the columns you used for combining. Then follow this up with an Extract Column Header node on those filtered columns. You can then use this row of column names to get the original column headers back into your new table.
Hopefully this should be quicker than looping 4000 times.
(I guess the delimiters/quotes should be different if the search or the replace pattern contain them.)
Thanks it works and its defintely quicker than looping