Re: An easy (I think) SQL question

On 29-Jun-2016 08:20 -0500, Bill Howie wrote:

[…] Is the TIMESTAMP_FORMAT function something that came along with
more current releases of the OS? We are, unfortunately, still on
V5R4.

AFaIK [and DeveloperWorks is effectively dead for me presently, such that I can not even try to lookup], the feature was added to, and only since, IBM i 6.1; and the "What's new for V6R1" of the IBM i 6.1 KnowledgeCenter confirms that addition:
[http://www.ibm.com/support/knowledgecenter/en/ssw_ibm_i_61/db2/rbafzwhatsnew.htm]
Note: knowing that feature was an addition to v6r1, does not imply that the function was not made available earlier, e.g. via PTFs to v5r4; but determining availability on an earlier release requires finding either a PTF documenting that support or instead\similarly finding some docs on ibm.com/developerworks

Per that scalar function being implemented as a UDF vs as a true built-in, there is IMO, little value over using that system-supplied TO_DATE versus using a user-created UDF such as the one I offered [or the variation explicitly defined as DETERMINISTIC and with error-handling as shown by Birgitta]. In fact, such a UDF that is specific to the task [esp. the version without error-handling for which surely inlining will be possible since a more recent release] rather than so conspicuously generic with what are offered as formatting capabilities, will almost surely perform [if even only negligibly] better.