Re: RE: st: Appending several excel data sets into one

["Michael Blasnik" <michael.blasnik@verizon.net>]

I think you miss the point.  Many Stata users, myself included, receive data 
they need to analyze from others in Excel format -- sometimes in many separate 
xls files.  These files often have data type (and format) issues -- at least 
when accessing them using default methods and the MS ODBC driver.  I don't want 
to launch Excel, edit the raw datasets (running macros, etc), save my new 
versions and then import these files into Stata.  Instead, I want to be able to 
work completely within Stata to get the data I need as it is provided in 
original format.  By using an all-Stata approach, I can produce a single do file 
that performs an entire analysis -- from raw data to final results.  This level 
of automation is critical to me so that I have a fully reproducible and 
documented analysis.  Just because Excel CAN do something doesn't mean that's a 
good approach to take.

Michael Blasnik

p.s. it appears that you have a sticky "?" key on your keyboard :)

----- Original Message ----- 
From: "Sergiy Radyakin" <serjradyakin@gmail.com>
To: <statalist@hsphsun2.harvard.edu>
Sent: Monday, September 17, 2007 8:52 AM
Subject: Re: RE: st: Appending several excel data sets into one



> Excel does not provide a facility to access sheet names?????????
> Download utilities????
> What are we talking about?
> ------------------------------------------------------------------------------------------
> Public Sub Liste()
<snip>> End Sub
> ------------------------------------------------------------------------------------------
> This will create an index sheet in an arbitrary Excel file.
> (No external utilities or black magic required).
> Will probably work on every Excel version from the late 1980s or so...
> Couple of lines can obviously be removed, but are left for clarity.
>
> As for the types of variables, you can simply declare cells of having
> cirtain type, e.g. numeric, or strings, or dates in exactly the same
> way you will do it in any other system. Since Stata supports only two
> types of variables, you do not see it very often explicitly, unless
> you want to create a new variable. (oh, yes, yes, there are date
> variables in Stata 10 too.)  However istead of directly specifiying
> the type, you rely on the "The DaVinci Code of the Missings":
>
> gen a = .
> gen b = ""
>
> so having seen this, any new-comer will be puzzled by the following line:
>
> gen c = "."
>
> What is the variable type? Is it assigned missing values ? How can I be sure?
>
>
> Excel will do a good job as long as it is programmed to do it. Just
> switch off all automatic "bells and whistles"  and tell it what you
> want to do in a program (same as you do in Stata).
>
> At least when it comes to programming, Excel is much superrior to
> Stata with its Yeti-like macros found in no other interpreters aside
> assembler.
>
> I have originally suggested to append the sheets in Excel, and after
> following the responces with StatTransfer etc. still think it is not a
> bad idea.
>
> Best regards,
>   Sergiy Radyakin
*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/