Re: CRTSQLRPGI parameter

On 11-Aug-2016 17:18 -0500, Brian Parkins wrote:

<<SNIP>> I would presume and expect that, for the lack of any
MARGINS capability, the effect will remain unchanged from that of
past releases, for the new default of *CALC; meaning, the qualified
name QTEMP/QSQLTEMP1, or the customized\changed default. And if any
version of the command is made available with the MARGINS support,
then anyone taking advantage, will probably /learn/ the effect is
similar to what is documented on those other two commands;
supposing, QSQLTEMP1 vs QSQLTEMP being the only difference in the
help text.?

[…] I did a little detective work and found that if the source file
has a record length > 112, then the TOSRCFILE will be named
QSQLTxxxxx, (e.g. QSQLT00113 if the source record length is 113).
Otherwise TOSRCFILE retains the name of QSQLTEMP1. Interesting!

That sure contradicts my expectation that the effects would await a MARGINS parameter keyword.

Anyone interested [in helping others] could pursue the help text getting updated, to reflect that the feature of *CALC for the To Source File (TOSRCFILE) has been implemented despite a lack of a MARGINS parameter being available in conjunction. Unsure how many people will be both using a RcdLen>112 and have some dependency coded on the defaulted name QSQLTEMP1 [in QTEMP], but I would hope that they\IBM have documented both the effect in the Memo To Users (MTU) for the first release the effect starts per the new default [IBM i 7.1 apparently], and that the preventive for ill-effects [for having such a dependency] would be to issue the following:

chgcmddft qsql/crtsqlrpgi 'tosrcfile(qtemp/qsqltemp1)'

A web search turns up nothing, so unlikely there is any MTU doc about that change. So probably if reporting the issue about the help text, then also the lack of doc in the MTU to warn of changes that were implemented at that release that might impact [even just a small subset of] users.