[R] Event History Data Recoding

[Don MacQueen]

The reshape() function would be used to reorganize your event history 
data set so that the Discussion and Agreement dates are one above the 
other in the same column, presumably with another variable to flag 
which is which (although you don't seem to need that).

Then, separately, you read your "information" data into R, reorganize 
it into the same structure as the reshaped event history, and then 
append it to the reshaped event history (using the rbind() function), 
and finally, sort.

That's my best guess given what you've shown. But there are puzzles, 
i.e., dates in the inputs that are not in the output, and dates in 
the output that are not in the input. Event=0 shows up in the output 
but not in the input. The "information" dataset doesn't have event 
numbers.

Reading the "information" file into R and reorganizing into columns 
with the correct structure for appending looks tricky.

-Don


At 5:18 PM +0200 4/23/10, Thomas Jensen wrote:
>Thanks Josh,
>
>But I am not sure if the reshape function can create new rows based on
>the meeting variable. For the second act there are three meetings, i.e.
>one meeting between discussion and agreement, and this should be entered
>as a separate row.
>
>Best, Thomas
>
>On Fri, 2010-04-23 at 07:45 -0700, Joshua Wiley wrote:
>>  If you are talking about changing from a "wide" format to "long" 
>>take a look at
>>
>>  ?reshape
>>
>>
>>  ---
>>  Josh
>>
>>  On Fri, Apr 23, 2010 at 7:09 AM, Thomas Jensen
>>  <thomas.jensen at eup.gess.ethz.ch> wrote:
>>  > Dear R list,
>>  >
>>  > I have an event history data set that is structured like this:
>>  >
>>  > Legislative act     Discussion     Agreement     Time     Event
>>  > Act1                2006-05-30     2006-06-19     20        1
>>  > Act2                2004-03-01     2004-06-14     105       1
>>  > .
>>  > .
>>  > .
>>  >
>>  > I have information on the meetings in the legislature between adoption
>>  > periods in a separate variable (the start and stop dates are included):
>>  >
>>  > Act1 meeting:2006-05-30, 2006-06-19.
>>  > Act2 meeting: 2004-03-22, 2004-04-26, 2004-06-14.
>>  >
>>  > I want to code this as a discrete event data set, so it should look like
>>  > this:
>>  >
>>  > Legislative act     Event     Meeting     Time Count
>>  > Act1                  0       2006-05-30      1
>>  > Act1                  1       2006-06-07      2
>>  > Act2                  0       2004-03-22      1
>>  > Act2                  0       2004-04-26      2
>>  > Act2                  1       2004-06-14      3
>>  >
>>  > Can anyone tell me if it is possible to do this recoding in R, or do I
>>  > have to do it by hand?
>>  >
>>  > Best, Thomas
>>  >
>>  > ______________________________________________
>>  > R-help at r-project.org mailing list
>>  > https://*stat.ethz.ch/mailman/listinfo/r-help
>>  > PLEASE do read the posting guide 
>>http://*www.*R-project.org/posting-guide.html
>>  > and provide commented, minimal, self-contained, reproducible code.
>>  >
>>
>>
>>
>
>______________________________________________
>R-help at r-project.org mailing list
>https://*stat.ethz.ch/mailman/listinfo/r-help
>PLEASE do read the posting guide http://*www.*R-project.org/posting-guide.html
>and provide commented, minimal, self-contained, reproducible code.


-- 
--------------------------------------
Don MacQueen
Environmental Protection Department
Lawrence Livermore National Laboratory
Livermore, CA, USA
925-423-1062