"Carl Daniel [VC++ MVP]" <cpdaniel_remove_this_and_nospam@mvps.org.nospam>
[quoted text, click to view] wrote
>I have a report that contains a single chart. That report works fine by
>itself, and as a subreport in many other contexts, but...
>
> When I use that report as a sub-report inside a table, passing all of the
> parameters to the sub-report based on values from the current row of the
> table...
>
> The chart is labelled correctly (I echo all parameters in the chart
> title), but the data that's charted is clearly wrong. The chart always
> has 3 series, but in any given table, only the first chart is correct.
> All the rest "randomly" (but reproducibly) either show all zeros for one
> or more (but never all) series, or show data that looks a lot like the
> rendered version of the data from the first chart (i.e. the data is the
> same magnitude and shape on the printed page even if the real magnitude is
> quite different).
>
> I'm sure this is some magic of RS trying to "help" me by "grouping" or
> "filtering" or in some other way "fixing" my data, since clearly I didn't
> want to simply see the results of my queries charted... but I'm at a loss
> to know what to do about it.
>
> Any ideas?
>
> I'm using RS 2005 SP2 on Server 2003 R2 SP2 x64.
PS: I can see in SQL Profiler that all of the correct queries are being
executed with the correct parameters - one outer query by the master report
and one query per row for the sub report, with parameters taken from the
results of the outer query. All queries are OLAP/MDX queries, in case that
makes any difference (I can't imagine how it could, but who knows).
PPS: Yes, I know this would be more efficient if I "simply" built a larger
query and used grouping in a list or table instead of a sub-report, but
that's not the point. The queries are complex, and composing them would be
challenging (at least!); plus, the number of rows from the outer query is
never large.
-cd
