Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, is already up and running.

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

st: SV: AW: SV: RE: Splitting a textvariable

From   "Tomas Lind" <>
To   <>
Subject   st: SV: AW: SV: RE: Splitting a textvariable
Date   Mon, 17 May 2010 08:33:27 +0200

Thanks Martin. Your code works wonderful. 

/ Tomas

-----Ursprungligt meddelande-----
[] För Martin Weiss
Skickat: den 12 maj 2010 18:02
Ämne: st: AW: SV: RE: Splitting a textvariable


So, Tomas, as Nick has already said, this is not trivial:



inp byte id str50 address
1	"Stockholm xroad 12 London yroad 10"
2	"London zroad 31"
3	"Helsinki rroad 1 Oslo sroad 123 Berlin troad 13"

split address, gen(newaddress)
drop address
reshape long newaddress, i(id) j(addnumber)
drop if mi(newaddress)

bys id (addnumber): gen byte runningnum=sum(real(newaddress[_n-1])!=.)+1

bys id runningnum: gen byte j=sum(runningnum[_n]==runningnum[_n-1])+1
drop addnumber

reshape wide newaddress, i(id runningnum) j(j)

gen address=newaddress1+ " "+ newaddress2+ " "+ newaddress3

drop newa*
l, sepby(id)


-----Ursprüngliche Nachricht-----
[] Im Auftrag von Tomas Lind
Gesendet: Mittwoch, 12. Mai 2010 17:24
Betreff: st: SV: RE: Splitting a textvariable

Thanks Nick and Bjarte for kind help and advice. 

I´ll going to learn more about regular expressions and -strpos()- and
hopefully get it to work.


-----Ursprungligt meddelande-----
[] För Nick Cox
Skickat: den 12 maj 2010 14:14
Ämne: st: RE: Splitting a textvariable

-inrange()- won't work as you want because the address is a string variable.
I can think of two strategies here:

1. Use -split- to split into string and numeric components; then put them
together again with a comma inserted. 

2. Use regular expression functions. 

3. Find the spaces with -strpos()- and stop when you reach the first space
preceded by a number. 

There is some programming needed in each case. 


Tomas Lind

I have (in principle) a dataset that looks like this with an id-variable and
a text-variable with addresses. Unfortunately there are often several
addresses in the same variable. I have to split these into one variable for
the first address, another for the second and so on.

id   	address
1	Stockholm xroad 12 London yroad10
2	London zroad 31
3	Helsinki rroad 1 Oslo sroad 123 Berlin troad 13

Each address ends with a number. My idea is to put in a parsing "," in the
empty space after the number (to be able to use -split parse-. Like this.

id   	address
1	Stockholm xroad,12 London yroad10
2	London zroad 31,
3	Helsinki rroad 1,Oslo sroad123,Berlin troad 13

One idea to do this is to use something like (if we rename the
address-variable to a shorter name, x).

replace x[i]=","  if x[`i' - 1]==inrange(0, x[`i'], 9) & x[`i']==" "

I´m not sure about how to work out the details to get this into work. Any
help is welcome to put the commas into place. Other ideas are also welcome.

*   For searches and help try:

*   For searches and help try:

*   For searches and help try:

*   For searches and help try:

© Copyright 1996–2016 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index