Allow me to suggest a feature for string replacement. Every once in a while, I have to replace numeric values with their textual representations, e.g. 1 by “Natural Person” and 2 by “Legal Person”. As it’s usually the case, both tables (the source table and the lookup table) reside side-by-side in the same data base.
However, the only thing KNIME will allow me at present is to:
- export the lookup tables into an unquoted CSV format, one file for each value to be replaced, and usually only after re-ordering the columns
- use the dictionary string replacer node on the numeric columns, but only after they have been converted to strings
Now this is very burdensome. While I am aware that the same operation can be performed in a single SQL query, it does make the query more complex and noticeably slower. But I sincerely prefer my queries to be simple, quick and easy to read - not mention that sometimes my source is a text file anyhow.
If you could come up with an easier concept of value replacement, i.e. using KNIME tables for lookup, and replacing int values by strings in one go, it would greatly enhance the utility of KNIME for commonplace ETL operations. Please give it a thought, as comments are very welcome.