Error importing from RMark

[bmiranda]

I am trying to import a model and results into MARK that were created in RMark, so that I can get the median c-hat estimate. I used export.MARK to create the *.Rinp file for import. MARK creates the dbf file to hold the results, but then hangs up and I receive the following error message, after which MARK closes:

ERROR CODE: 11[]
Subsystem: BASE
Description: Dynamic Memory Low
Call Stack:
MYTOPAPPWIN: APPENDOUTPUT (Line: 100)
MYTOPAPPWIN: FILENEWRMARK (Line: 298)
APP: START (Line: 855)
START (Line 6)

Does this error really mean that the computer does not have enough memory? I am running on a 64-bit Vista machine with 16GB RAM. I believe MARK is a 32-bit program so it cannot access all of that memory, so running out of memory probably is possible, and the model is fairly complex. I have seen a few other posts that reference this same error code, but no responses to any of them. Can anyone confirm that this error does indicate a memory issue? Does anyone have any suggestions as a work around? Can the import process be done in multiple steps for example? Can I import just the model without the results?

I will appreciate any advice on how I might be able to get around this issue. Thanks in advance for any help!

-Brian

[jlaake]

I think the only one that will be able to help you is Gary. I've not had this issue. Can you give me some measure of what you mean by a complex model. What model? How many occasions and how many capture histories. You can use the .inp file it creates and use the standard way of creating the project in MARK by interactively answering the questions. The .Rinp file is a text file and you can see what it sets in it. Then once you get the project set up, you can import the analysis (the tmp files) with Output/Append on menu. My guess is that is where it will choke. You seem to have sufficient memory. --jeff

[bmiranda]

Jeff, thanks for the advice. I will try the manual method of creating the project in MARK. Since you asked, here are the stats on my "complex" model: 5 capture occasions; 134694 capture histories; 4 variables defining groups, combining into 87 unique groups; and 4 individual covariates.

Here are my model parameters:
p.dot=list(formula=~1,fixed=1)
Phi.complex=list(formula=~SC + ET + PFT + exp(DISTEDGE) + SC:exp(DISTEDGE) + PFT:exp(DISTEDGE) + PC1 + PC2 + PC3 + ZONE_F)
where SC, ET, PFT and ZONE_F are groups; and DISTEDGE, PC1, PC2, and PC3 are individual covariates.

If there is any additional information I can provide that will help diagnose this issue let me know.

Thanks,
-Brian

[bmiranda]

I am having a new problem with importing models from RMark into MARK. I have exported the models using export.MARK, and when I try to import them into MARK, I get an error: "Improper data type specified." If I try to import the model a second time, without closing the MARK window, I get a different error:
Error Code: 2 []
Subsystem: BASE
Function: ARRAYGET
Argument: 6
CallStack: MYTOPAPPWIN:FILENEWRMARK (Line:198)
APP:START (Line:855)
START (Line:6)

This is frustrating me because I have successfully gotten at least 1 model to import into MARK after resolving the issue I had in the previous posts. That model still imports fine, after deleting the residual file which was causing my earlier problems. I have examined the .Rinp files for the old model and the new ones, and I don't see any clear differences. I have identified a couple differences between the old model that works and the new models that don't:
1) The old model was the only model run in that workspace (mark001) and the new models have multiple models in the workspace. The models I am trying to now import are mark002 and mark003. Each time I am exporting a single model, not the marklist of results.
2) The old model contained individual covariates and the new models do not. This is the only obvious difference between the .Rinp files.

As far as I can tell, none of these differences should change whether the exported model can be imported to MARK. If anyone has any thoughts on what's going on here I would love to hear them.

Thanks,
Brian