Re: st: Translate Google Trends date system into Stata date system
Date
Wed, 01 May 2013 12:41:12 -0400
Or, for maximum concision (and obfuscation):
forval i=1/2 {
generate date`i' = date(substr(googledate, 12-11*`i' ,10),"YMD")
format date`i' %td
}
On 5/1/2013 12:21 PM, Sergiy Radyakin wrote:
tempvar d1 d2
generate `d1'=substr(googledate,1,10)
generate `d2'=substr(googledate,-10,.)
generate date1=date(`d1',"YMD")
generate date2=date(`d2',"YMD")
format date1 %td
format date2 %td
list
On Tue, Apr 30, 2013 at 3:13 PM, Nick Cox <njcoxstata@gmail.com> wrote:
That's fine.
But in order that no-one gets confused:
1. -parse()- here is the name of an option.
2. -parse- is a quite distinct and out-of-date command.
The connection is only that I had parsing in mind when I wrote the
original version of -split-, itself a rewrite of -strparse- by Michael
Blasnik and myself, still visible on SSC.
Nick
njcoxstata@gmail.com
On 30 April 2013 20:05, Stephen Cranney <scranney@sas.upenn.edu> wrote:
The split and -parse- commands worked perfectly for my purposes. Thank you.
On Tue, Apr 30, 2013 at 2:38 PM, Phil Schumm <pschumm@uchicago.edu> wrote:
On Apr 30, 2013, at 1:27 PM, Stephen Cranney wrote:
I'm trying to get long series of dates I downloaded from Google Trends to function in Stata. Unfortunately, Google Trends's way of doing dates makes this difficult. An example value from a single observation:
2004-01-04 - 2004-01-10
This represents the week of January 04, 2004 to January 10, 2010. However, Stata is (understandably) reading this as a string variable. Is there a way to take variables like this and break them into a beginning date and ending date or something similar?
