time saver, how to organize your occasion history

[hpdutra]

Chapter 2 of gentle introduction says that it is better to organize your encounter history as
1000 2 1;
1011 1 3;

instead of
1000 1 0;
1000 1 0;
1000 0 1;
1011 1 0;
1011 0 1;
1011 0 1;
1011 0 1;

especially for large data sets. Well in my case I already had a huge encounter history for each individual, and I did not want to create a huge formula in excel.
If you know how to use R here is an easy way of doing it using the package "reshape"

http://www.nabble.com/RESHAPE-package-q ... #a18794249

You still have to export to excel and remove the ID column and transform it into INP, but boy this saves time!

[cooch]

Chapter 2 of gentle introduction says that it is better to organize your encounter history as
1000 2 1;
1011 1 3;

instead of
1000 1 0;
1000 1 0;
1000 0 1;
1011 1 0;
1011 0 1;
1011 0 1;
1011 0 1;

especially for large data sets.

Not quite - the summarized format has some efficiencies - especially for large datasets, but precludes use of individual covariates. In most cases, there is little lost by using individual encounter histories (rather than summarizing over histories), but doing such a summarization is easy enough...

Well in my case I already had a huge encounter history for each individual, and I did not want to create a huge formula in excel.
If you know how to use R here is an easy way of doing it using the package "reshape"

http://www.nabble.com/RESHAPE-package-q ... #a18794249

You still have to export to excel and remove the ID column and transform it into INP, but boy this saves time!

Useful tip. You can do the same in SAS in a couple of lines (actually, one call to PROC MEANS). In fact, any of the semi-standard data manipulation tools can handle this sort of thing easily, which is why MARK doesn't have its own database capabilities - they would be redundant to what is already sitting on the desks of most folks.

[hpdutra]

Hi Evan
I actually know how to do that in SAS, but I don't have that at home. So figuring that out in R was great. I am going to quote you "you can't beat the price of R"

[dhewitt]

There was some confusion in the thread over at R-help. For folks looking to use this handy shortcut, I posted a clarification. It's simple to do with the 'recast' wrapper that Hadley provides. Note that this particular usage won't create the proper format for more than 2 groups. And, as Evan pointed out, you'll often want covariates so the condensed format won't be the right way to go. But, here it is:

http://www.nabble.com/Re%3A-RE-SHAPE-pa ... 14133.html