|
PureBytes Links
Trading Reference Links
|
Jim,
By "under similar circumstances", I meant to imply an apparent
similarity between error messages which I had received and error
messages which you were reporting; I did *NOT* mean to imply that I was
trying to Paste In a .dat file with no associated .idx file.
To be more specific: I was attempting to paste in a BMI/S&P data file
from Omega's Web site which turned out to be corrupted (they replaced it
the next day).
The fact that you received no CTREE error message is bothersome although
my gut feeling tells me that this is not a significant issue. (Omega may
not have included that message until Release 4.0, for example...an
assumption which may or may not be true, of course).
The point I'm trying to get to is this: in spite of the fact that you
are not reporting a CTREE error message, everything else you are
reporting points toward a corrupted set of ISAM pointers. Assuming that
everything was ok up until your last Paste In operation, I would then
maintain my original suggestion that the root cause of your problem is
very likely "that you were inadvertently pasting in data contained in a
corrupted data file [which] then corrupted your index file".
If that is the case, then the methodology which I had previously
outlined will rebuild your ISAM pointers...something which needs to be
done, by the way, *before* you attempt to "[copy] out some data from TS,
and [remove] the index file, and [try] to paste-in that data file".
Without first rebuilding your ISAM pointers, you will defnitely get what
you got: "the same error msg box".
You were right, in other words, to "immediately [distrust your] .dat
file"...it was just a matter of first rebuilding those ISAM pointers
*before* testing your supposition.
If my line of reasoning is correct, your primary problem then becomes
one of fixing the corrupted data file you were attempting to Paste In at
the time your troubles began (after you re-build your ISAM pointers, of
course). Whether or not "the BMI date fiasco" is in any way related to
that problem remains an obvious unknown (but an obvious one worth
pursuing).
Dave
Jim Osborn wrote:
>
> "David L. Miller" <dlmiller@xxxxxxxxxxxxx> responds:
> >
> >Jim,
> >
> >Did you not also get a CTREE error message while your Off-line Server
> >was up (something to the effect of "Omega Server could not open Target
> >File")? If so, chances are that you were inadvertently pasting in data
> >contained in a corrupted data file. This then corrupted your index file
> >(ISAM = Indexed Sequential Access Method). In general, CTREE errors
> >imply corrupted index files.
>
> Dave,
> Interestingly, there were NO error messages in the server log. All it
> said was the usual "Rebuilding ISAM files" but it never put any other
> error msg in the log. At the end of the evening, the server log had
> a string of "Rebuilding..." messages one right after the other, but
> never any "CTREE" errors.
>
> >Here's what I did (that worked) under similar circumstances about a year
> >age (running TS 4.0):
> >
> >1. Delete: 40BONN5.IDX; 40BONN5.DAT; 40BONN.IDX
> >2. Bring up Off-line Server
> >3. Exit Off-line Server
> >4. Bring up On-line Server
>
> Now that the weekend is here, I'm about to try a bigger hammer.
> Since I was getting a message that seemed to imply that the basic
> CTREE engine was failing, I was reluctant to delete my main-data .idx
> file, which, of course, is the normal recourse for all data problems.
>
> When you say "similar circumstances" do you really mean you were trying
> to paste in a .dat file with no associated .idx file? Will TS 4
> play that game?
>
> Thanks for the suggestions. I'll let everyone know how it goes...
>
> Jim
>
> Jim Osborn <jimo@xxxxxxxxxx> writes:
> >> When I try to create a new ISAM data index file, I get the error msg:
> >>
> >> "OpenIsamFile failed case 1"
> >>
> >> in a msg box.
> >> The only choice is to acknowledge the box and abort the rebuild.
> >>
> >> I'm getting this error msg during a paste-in operation. I don't
> >> know if I'd get the same error if I remove the .idx file from my
> >> main data file, but I'm not about to risk a test at this point.
> >>
> >> Is anyone familiar with this part of TradeStation? What could
> >> have come undone? As far as I know, nothing has changed in my
> >> filesystem, but unless TS has a timebomb that says OpenIsamFile
> >> will stop working as of Sept 10, 1998, SOMETHING must have got
> >> lost, some dll, or something?
> >>
> >> I find it hard to believe that the BMI date fiasco could have
> >> broken something, but that's the biggest disturbance I've had
> >> recently. I haven't even had a GPF for months. I am getting
> >> a bit short on disc space. Should rebuilding an ISAM index
> >> take massive amounts of disc? Copying out doesn't seem to.
> >>
> >> Needless to say, this failure is a disaster of major proportions.
> >> As many of you know, I'm still using TS 3.5, primarily because,
> >> until last night, I could import custom data into its tick database.
> >> This ability allows, for example, backup data from sources like
> >> FutureSource, and most importantly for me, allows creation of
> >> real-time "continuous" contracts, so that at S+P rollover, I've
> >> got a meaningful chart of the new contract, using the adjusted
> >> old contract. I create a TS-style .dat file from whatever data
> >> I'm starting with, then put it in the paste-in directory and let
> >> TS "rebuild the corrupted index file." Until last night.
> >>
> >> Of course, I immediately distrusted my .dat file. But when I
> >> copied out some data from TS, and removed the index file, and
> >> tried to paste-in that data file, I got the same error msg box.
> >> So I'm pretty confident that something has broken in TS.
> >>
> >> I had hoped to share the rollover data I created last night, as
> >> I did last rollover, but I don't know if it'll work for anyone
> >> without that .idx file. In the meantime, I'll put the two files
> >> I DO have, the no-suffix file, and the .dat file, on my ftp site
> >> later today, and if anyone cares to try pasting-in with those
> >> (I'm pretty sure you'll need TS 3.5, not TS 4.0) files, and if
> >> it works, maybe they'd be so kind as to send me the .idx file
> >> so I can use it too.
> >>
> >> And if someone can help me get my own paste-in function working
> >> again, I'll share the format of the .dat file, so you can write
> >> your own data generator.
> >>
> >> Jim
> >
|