Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: RE: AW: RE: AW: Categorising dates


From   Nick Cox <[email protected]>
To   "'[email protected]'" <[email protected]>
Subject   st: RE: AW: RE: AW: Categorising dates
Date   Tue, 24 Aug 2010 16:03:08 +0100

Just a mild protest, as signalled. I stand by my arguments. Sure, the efficiency gain is not detectable at n = 50.  

Nick 
[email protected] 

Martin Weiss

" gen mydays = string(mydates, "%tdMonth") 

which replaces a call to an .ado which is dozens of lines long with a single
line of code with exactly the same effect."


Both calls are a single line long:

*************
gen mymonth1 = string(mydates, "%tdMonth") 
tostring mydates, gen(mymonth2) format(%tdMonth) force
*************

And both work out at "0.00" seconds on my computer (-set rmsg on- to see for
yourself), so the benefit has got to be so slight not even Stata notices...


"Respecting the problem"

What is this heading supposed to mean? I gave Sara a solution that is
intelligible when -list-ed to the Results window. Most other solutions
require you to label afterwards using techniques as in your very own
http://www.stata-journal.com/sjpdf.html?articlenum=pr0013 (What is "3"
again, as in -di in r dow(date("23 Sep 09", "DM20Y"))- ? Solution: A
Wednesday...)


Generally, everything depends on what Sara wants to use the results for. In
the absence of this information, we can only guess...

Nick Cox

As the putative author of -tostring-, I must protest mildly at this use of
-tostring-, on two quite different grounds. 

1. Style and efficiency
=======================

If you are working with a numeric variable, are inclined to allow force, and
wish only to generate a single string variable, you can and should get there
directly with e.g. 

gen mydays = string(mydates, "%tdMonth") 

which replaces a call to an .ado which is dozens of lines long with a single
line of code with exactly the same effect. 

-tostring- is a convenience command which is, literally, convenient when (a)
you have two or more variables and/or (b) a desire to be prudent because you
are worried about loss of information in conversion. If neither applies,
calling up -tostring- is unnecessary. 

2. Respecting the problem
=========================

For problems like Sara's the user is almost always better off with numeric
date variables assigned appropriate date formats. 

Nick 
[email protected] 

Martin Weiss

clear*

//generate data
set obs 50
gen mydates=date("23 Sep 09", "DM20Y")+_n-26
format mydates %tdMon_dd,_CCYY

//Get day of week
tostring mydates, gen(mydays) format(%td_Dayname) force

//Get month
tostring mydates, gen(mymonth) format(%tdMonth) force


//see result
l, noo
*************

sara khan

I have a list of daily dates inthe format, for example, 23 Sep 09, and
need to create two variables. One is to categorise the days into
weekly data (so week commencing on a Monday). The second is to create
a variable cataegorsing the daily data into monthly data.

I would be grateful for advice on how to do this.


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/


© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index