Date: Mon, 18 Apr 2011 13:45:44 -0400
Reply-To: Quentin McMullen <qmcmullen.sas@GMAIL.COM>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Quentin McMullen <qmcmullen.sas@GMAIL.COM>
Subject: Re: tagsets.excelxp multiple files when HTML is on
In-Reply-To: <4DAC75EA.3080200@gmail.com>
Content-Type: text/plain; charset=ISO-8859-1
Well that sounds pretty scary. In the past some little birdies have chirped
about these sorts of issues when tagsets came up. Maybe one will offer
additional explanation. Congrats on getting it solved (if not completely
understood...).
--Q.
On Mon, Apr 18, 2011 at 1:33 PM, Ben HErman <baherman@gmail.com> wrote:
> Q,
> This was one that would make you pull you hair out. I could easily
> replicate the problem (just startup SAS and run the code -- any
> tagset.excelxp code -- even new code I typed in). It was so insidious
> because it turned out my final conjecture ended up being correct. Something
> in sasuser.regstry must have gotten mangled. I backed it up, restarted SAS,
> (which recreates it with it's defaults), and the problem magically
> disappeared. If I hadn't opened the second session (to do basically exactly
> what you suggested: add lines of code until I encountered the error), I
> never would have discovered the problem, because only in the second session
> with the default registry did the problem disappear, and thus become
> obvious.
>
> Thanks for all the help and suggestions. Mucho Gracias (Cinco del Mayo is
> quickly approaching!)
>
> Ben
>
> On 4/18/2011 1:23 PM, Quentin McMullen wrote:
>
>> On Mon, 18 Apr 2011 11:28:39 -0400, Ben Herman<baherman@GMAIL.COM>
>> wrote:
>>
>>
>> No Joy here! You ever have one of those days where nothing makes sense?
>>>
>>> I appologize that my sample code was muddled (I accidentally dumped the
>>> options when I shortened it up for posting). I had originally included
>>> the sheetname. So here's the kicker. I decided to do a little more
>>> trouble shooting. Opened up SAS ran my code -- got the error. Opened up
>>> a
>>> second session, copied the code in sections -- No error. Opened up a
>>> third session, copied all the code from the first session into the third,
>>> ran with no error. Apparently I have poltergeists.
>>>
>>
>> Hey Ben,
>>
>> I had a similar poltergeist a couple weeks ago. I would keep trying to
>> replicate the problem, but it seemed to happen randomly. After I slept on
>> the problem I woke up the next day and was able to replicate it
>> consistently. And it turned out in my case (and maybe yours?) I had
>> accidentally forgotten an ods tagsets.ExcelXP close statement, so I had
>> multiple destinations open. And this caused tagsets.ExcelXP to write bad
>> xml. (Sounds suspiciously like your first post, where you had an html
>> destination open...) I've also gotten errors when Excel tries to open a
>> file with the title/footnote is apparenlty too long (for Excel to read,
>> not too long for SAS to write). But once you get through the hour or 2
>> (or 8) of thinking "well, this statement can't be the cause, so I'll
>> delete it from my code", and suddenly replicate the problem, it
>> becomes "obvious".
>>
>> So I would close SAS, go for a walk by that pretty waterfront outside your
>> office, and come back and try again with a new session. Then repeat.
>>
>> --Q.
>>
>
>
| 
