bys: outreg2 is returning error message: invalid 'onebyone' r(198)

[salma ktat]

Hi everyone,
I just installed stata13 in a new computer (I used to use Stata in my old computer and everything was good). So now I'm having "weird" messages such as unrecognized outreg2 (so I installed it, but I was thinking that stata13 already contain outreg2!). Anyway, the new error message I'm having is that it didn't recognize the bys option of outreg2. So, I checked the manual and it's there, I mean that outreg2 has a bys option. So, can enyone make me understand why stata is returning these errors and how to correct for the last one.
Thanks.

[Attaullah Shah]

bys option can be used only in a specific combination of commands, what did you type and what was the result?

[Nick Cox]

Part of this is easy: outreg2 is user-written (SSC) so doesn't come bundled with official Stata. You need to install it when setting up a new computer.

As for the rest: outreg2 supports the bysort: prefix command (which is not an option, but as said a prefix command). So you are telling us that you got an error but not showing us your exact code. I think you need to show exact code, so that users of outreg2 (not me) have a chance to advise.

Alternatively, contact the program author directly: he is not active on Statalist.

[salma ktat]

Thanks Attaullah,
I typed the following

bys Group: outreg2 using mydoc, append title ("Table : Summary statistics by groups of companies") sum(log)

Knowing that Group is a variable that takes two values 0 and 1 to distinguish between two groups of observations.
The stata message is the following:
invalid 'onebyone'
r(198);

And I checked for the outreg2, normally the bys prefix is allowed!
Thanks again.

[Nick Cox]

Does

 
onebyone

correspond to anything in your dataset?

[salma ktat]

Thanks Nick.
No, it does not!

[Attaullah Shah]

bys Group: outreg2 using mydoc.doc, append title ("Table : Summary statistics by groups of companies")  side replace : reg y x1 x2 x3

Try the above code, replacing y and x with your variables.

For summary statistics, you can use

bys Group: outreg2 using mydoc.doc, append title ("Table : Summary statistics by groups of companies") side replace : sum y x1 x2 x3

[Nick Cox]

As said, I don't use outreg2 but it's easy enough to search the code and to find that onebyone is an option used by a program defined within
outreg2.

So an internal call within the program is triggering a syntax error. Who knows otherwise why that could be. I think you need to contact the program author, Roy Wada, and send him a reproducible example of your error.

[salma ktat]

bys Group: outreg2 using mydoc.doc, append title ("Table : Summary statistics by groups of companies") side replace : reg y x1 x2 x3

Try the above code, replacing y and x with your variables.

For summary statistics, you can use

bys Group: outreg2 using mydoc.doc, append title ("Table : Summary statistics by groups of companies") side replace : sum y x1 x2 x3

I tried your solution and I got the not sorted error message even though the data is sorted by the id variables!!!

[Attaullah Shah]

You have to sort the data by variable that is used in the outreg2 command. This error message implies that the original problem has been solved.

[Attaullah Shah]

And if the problem stays, try this alternative

tabstat y x1 x2 x3, statistics( mean count max sd min ) by(group) columns(variables)

Copy as table and paste in Excel.

[Vaughan Connolly]

Although this is an old thread, it still seems relevant to the current version of outreg2 running in Stata 16. When replicating Example 16 (Group summary table) from the help guide, using the help file code, I receive the error message given above (invalid 'onebyone'). After uninstalling and then reinstalling both outreg and outreg2, the error persisted.

However there is a solution of sorts. When removing options one by one, I found that using the drop() option was causing the error (in this case). While drop() wasn't used in the original post above, if this problem reoccurs, I recommend trying this command with the smallest number of options that you can manage with.