I am currently reading a csv file which results in a table with columns that are not labeled and data that falls under the wrong columns. Is the CSV file reader maybe faulty here? The reason I mention this - it seems that the file parsing process is using the delimiter first to separate columns and then applies the quote character filter to the data? This means that: a file that has the format “AA”,“BB,BB”, “CC” - will split into AA BB BB and CC (i.e. 4 columns) - where it should only split into [AA] [BB,BB] [CC] - three columns. The reason I show “BB,BB” is because this column contains data that also contains commas, but is between quotes. Excel seems to read this file just fine, but somehow I cannot get KNIME to read the file correctly. Any advice?

Typically the File Reader ( File Reader — NodePit ) instead of CSV Reader gives you more control about the structure of imported tables. Also I used the R package readr to handle ‘messier’ CSV files. [image] Error with File Reader node while running in a loop KNIME Analytics Platform I had a similar problem where there were some ‘dirty’ CSV files out of a large number of files and they were too large to just ‘edit’ them in Notepad++ (and that also was not a very nice process). I built a workflow using the R package readr, which seems to me more tolerant towards problematic CSV files and gives you some control how to handle mismatching quotes and so on. You might have to toy around with the settings of the read_delim function. [46] kn_example_import_csv_r_readr.knwf (52.…

CSV Reader Bug

KNIME Analytics Platform

mlauber71 September 28, 2018, 8:49pm 2

Typically the File Reader (File Reader — NodePit) instead of CSV Reader gives you more control about the structure of imported tables.

Also I used the R package readr to handle ‘messier’ CSV files.

2 Likes