Fred Wolfe <[email protected]> writes,
> The graph editor is quite nice and very useful. However, I have run
> into a problem. I was making an hbar graph, made many modifications,
> and then modified blabel to add a % sign (what many of us had asked
> for on the list). The editor accepted the % sign. I saved the graph,
> but when I tried to use it I received this message:
>
> plotregion1.barlabels[8].text invalid name
> error (198) reading file
> Error in file, skipping lines:
> .plotregion1.barlabels[8].text.Arrpush 3.8%
> <EndItem>
> graph1.gversion invalid name
> [...]
This is a bug and we will get it fixed. Fred will not have to live without
percent signs in bar labels.
Until then, if Fred will send me his graph, I will edit it and send it back
with the offending percent signs removed, so that he won't lose all the rest
of his work on the graph.
(FYI, all I am intending to do is edit out the currently offending characters
as suggested by Tom Steichen <[email protected]>.)
I believe that the graph file is fine and the problem is in reading the graph
back, so once the fix has been made Fred should be able to read his original
edited graph, complete with the added percent signs.
After some admittedly incomplete testing, I have found that percent signs work
everywhere I have tried except in bar labels, that is to say they work in
titles, axis titles, etc. I would like to hear, either privately or on list,
if anyone finds cases other than bar labels where they do not work.
I would very much like to have a copy of Fred's graph, even if he doesn't want
me to edit it. I am unable to reproduce the error message Fred is getting.
I am seeing a different problem; some edits, including most "Object-specific
properties" changes, to bar labels are simply ignored when the graph is read
back. I do not believe fixing this will fix what Fred is seeing with the
percent signs.
Sergiy Radyakin <[email protected]> suggested that Fred edit the official
file _fr_runlog.ado. Sergiy was careful to suggest making a backup first.
Regardless, I would not edit this file. If that is where the problem lies,
and I do not believe it is, let StataCorp make that fix.
Sergiy notes that several strings are single quoted in _fr_runlog.ado, rather
than the safer compound quoted. I believe all such cases are for strings that
have been fetched as tokens, which by definition do not need compound quotes.
Sergiy noted that this single-quoting has caused him some problems in the
past. That surprises me. If Sergiy will send me some examples of this
privately, I'll look into them and report back to the list on anything
interesting.
-- Vince
[email protected]
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
