Error with large number of individual covariates

[jlaake]

strwrap in R could be used to do something like this. It may have to be used in a loop if you want to include comments as Evan showed. strwrap has arguments initial and prefix to add text at the beginning of lines.

[markmiller]

Thank you, Drs. White, Cooch and Laake, for the help. I will try breaking the lines, which are much longer with some of my real data because most of the covariates are not zero for some groups.

Evan's symbol /* 11 */, I think is just the line or record number (here the site number) and is ignored and MARK would identify the end of a record by the semi-colon.

Curiously, I ran some models this fall in RMark using all 10 individual covariates in a model without seeing this error. It was only when I tried running models for a single group directly in MARK that I first encountered the error. Although, when running models containing all 10 covariates I used a Linux machine.

I will probably post again after attempting the line-break solution with the real data.

[cooch]

Evan's symbol /* 11 */, I think is just the line or record number (here the site number) and is ignored and MARK would identify the end of a record by the semi-colon.

Correct -- MARK ignores comments regardless of where they occur. Gary suggested putting comments at the start of every line, simply as a way to keep track of things. I agree -- useful suggestions.

Curiously, I ran some models this fall in RMark using all 10 individual covariates in a model without seeing this error. It was only when I tried running models for a single group directly in MARK that I first encountered the error. Although, when running models containing all 10 covariates I used a Linux machine.

Which may be the explanation -- the way gfortran handles long input lines differs between Windows and Linux. It would not surprise me that running things under Linux eliminated some/all of the problem in the first place.

Gary and I are working on an updated Linux version of MARK, which might help. Stay tuned...