[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: How to tell a macro to use previous global value of that loop, akin to [_n-1] for columns

From   "Jared Fotis" <[email protected]>
To   [email protected]
Subject   Re: st: How to tell a macro to use previous global value of that loop, akin to [_n-1] for columns
Date   Sun, 17 Aug 2008 15:48:25 +0200


Thank you for the timely response on a sunday.  Indeed that is but one
line code for a long macro that i thought it better not to post in its
entirety.  You are correct that i have vars LHS_NMCE_1 through _10,
however these variables are equal to 1 2 3 or 4 representing the group
of that firm for that year, which can be continuously changing.
Therefore, the ==5 and ==1 in the example are meant to be hard code as
the 1 represents the associated group for that firm in the year before
that firm exited (LHS_NMCE_* value in the previous `X' global loop)
while the 5 signals the year after a group 1 firm disappears
(replacing the blank LHS_NMCE_* for that specific `X' loop).  This
line of code is then repeated for groups 2 3 and 4 by replacing them
with 6 7 and 8, respectively, in the first year after the firm exited.

This is all done to see the rate at which firms from different groups
drop out of the data by comparing the number of 5's in that year to
the total number of 1's in the year before (while dropping out firms
that have already exited, via a series of "save" and "use" commands in
the macro, so that they are not double counted).  As my question only
pertains to the way in which you tell stata to use the value of the
previous loop (or even just the variable with a name of that loop
value minus 1, if forgetting about the values in the global helps), I
thought it unnecessary to go into such detail.  The reference to
[_n-1], which refers to the value of the previous row, is that I am
hoping there is an equal expression to instruct stata to refer to the
previous value of the global loop it is then running.  Does such a
shortcut exists, and if not, has anyone found a solution in their own


*   For searches and help try:

© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index