Github user kaspersorensen commented on the issue:
https://github.com/apache/metamodel/pull/171
Hmm I "fixed" substring, at least ensured that it's returning the right
results, similar to Java's `String.substring` method. But it looks like you're
right with regards to the interpretation of SUBSTRING parameters in SQL. At
least I looked at PostgreSQL, SQL Server and MySQL, and they all have variances
of parameters, but one syntax in common: `SUBSTRING(str, start, count)` where
`start` is 1 based start index and `count` is "number of characters", not end
index. I guess we should comply with that?
---
