split function does not work in Cloudera Impala

Question

I keep getting an AnalysisException that says "split unknown" when I try to use the split function in Cloudera Impala. It seems to be a valid function listed on the built-in functions page. For reference, I'm using Hue to interact with Impala. Does anyone know the cause to this issue?

Solution 1

The following works in Impala:

split_part(string source, string delimiter, bigint n)

Documentation is here: https://www.cloudera.com/documentation/enterprise/5-9-x/topics/impala_string_functions.html

Solution 2

I asked the same question in the Google group and I was able to get an official reply. Basically, it's a misprint in the docs. split() is not supported because it would return an array and complex types are not yet supported by Impala.

Comments

[Markus]

Hmm while I appreciate your input, this does not answer the question at all + split_part only returns the nth element of the split.