I'd like to recode some of the values in one column to be an other value. e.g., a value of "999" should be replaced with "" (nothing), because the initial dataset had the 999-code as missing value. The other values should be left unchanged.
The column is of the integer type.
I haven't found any node that could do that operation (which I thought should be a simple one, because I'm used to working with SPSS in which this is accomplished very easy)
Just use the java snippet node and select to replace your value
Only problem, you have to accept any real number as the missing value.
Second workaround, resulting in real missing values. Use the row spliter to split the rows into the ones with 999 and rest.
Delete with the column filter your column in the 999 part
concatenate the modfied 999 with the unmodified rest. Done
May be more simple way is to use Number To String or Rename and String Replacer nodes. In more complex cases I used Rule Engine node.
Hi Iris, hi Max,
thanks for your replies. i'll try both ways. also thought of the number-to-string, replace and back again to number-approach, but i hoped that there could be a node that is just doing what i wanted...
Yepp, we are currently missing a node that does it all in one step. A node called "Insert Missing Values" or so.
Luckily, you can work around it but we put it on the list anyway.