Need help with Reshape!!! :(

[Alyssa Buchardt]

Hi guys,


I have data that looks like this:

ymd	year	month	day	time	output
200011	2000	1	1	12:00	0.123456
200011	2000	1	1	13:00	0.987654
200011	2000	1	1	14:00	0.456123
200012	2000	1	2	12:00	0.789123
200112	2001	1	2	13:00	0.963852
200112	2001	1	2	14:00	0.852852

And I need it to look like this:

ymd	year	month	day	12:00	13:00	14:00
200011	2000	1	1	0.123456	0.987654	0.456123
200012	2000	1	2	0.789123	0.963852	0.852852

When I use the following command:

"reshape wide output, i(ymd) j(time) string"

I get the error:

"output12:00 invalid variable name"



Is there any way I can overcome this?

Thanks,
A

[Jorrit Gosens]

Drop the colon ':' from the time variable. This character isn't allowed in variable name. At least that is the case in Stata 13.1.

[Oded Mcdossi]

Hello and welcome to the forum,
Please read the FAQ especially 6.
You can try the following code to avoid invalid characters in variable names.

​forval i=12/14 {
    replace time ="a_`i'" if time=="`i':00"
}
reshape wide output , i(ymd)j(time, string)

[Nick Cox]

Despite good advice you have received, I doubt that a reshape wide is really a good idea for your kind of data. It usually makes analyses more difficult, not easier.

P.S. I want to underline strongly our preference for full real names here, as pointed out by Oded. Please fix that.

[Alyssa Buchardt]

Thanks for the advice guys, and have requested a name change too.

Nick Cox, I know this is not ideal for analysis, I'm actually just using stata to manipulate the data into this format so I can export it to excel again for the purposes of updating an existing file used for analysis is linked to heaps of excel workbooks!