st: Re: tosql datatypes

[Kit Baum <baum@bc.edu>]

I am not sure what convenient syntax could be used to achieve this.  
For instance,

webuse auto
gen long priceyen=100*price

will create a long int, which is probably what you want large  
integers to be represented as in SQL. But if you just use generate  
without specifying datatype, it would make it a float. In this  
dataset, you cannot just treat floats as longs, as -headroom- would  
then be mangled. I suppose you could have as many options as there  
are SQL datatypes, and allow something like

tosql, table(junk) cre int(...)  bigint(...) float(....)  double( ...)

providing for varlists with wildcards in each case. Would that make  
sense? Variables not listed in any of the varlists would take on  
their Stata datatypes.


Kit Baum, Boston College Economics and DIW Berlin
http://ideas.repec.org/e/pba1.html
An Introduction to Modern Econometrics Using Stata:
http://www.stata-press.com/books/imeus.html


On Jun 13, 2007, at 2:33 AM, Henri wrote:


> tosql writes datatypes for variables when creating a table creation  
> script.
> But it uses the Stata datatypes for this, not the sql datatypes. I  
> have a
> column of integers, some around 36000. Stata treats this as float,  
> but this
> has a different meaning in sql. This is inconvenient as you will need
> conversions or casts in the program using the sql database to make  
> sure that
> computations won't fail on the datatype.
>
> Is there a way to improve on this? For instance via an interface to  
> tosql
> that allows for overruling the type of certain columns (or all at  
> once).
*
*   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/