Gabor Grothendieck <[email protected]> [Sat, Jun 19, 2010 at 11:19:35AM
CEST]:
> On Sat, Jun 19, 2010 at 4:56 AM, Johannes Huesing <[email protected]>
> wrote:
> > A rule would be a data structure containing the ID of the rule, the
> > rule in human readable language, an expression evaluating variables
> > within the environment of the appropriate data frame (and resolving to
> > a logical vector), possibly the data frame itself, and the query
> > message (possibly as an sprintf expression). The data structure may be
> > an S4 object or a list. In our current workflow, we manage the
> > validation rules using a spreadsheet, and import them into a
> > competitor's analysis software.
> >
> > I could pass
> >
> > function(df) with(df, (vsstresn < 30 | vsstresn > 130) & vstestcd == "HR")
>
> Another possibility would be to use an sql statement or fragment so if
> you passed this sql fragment:
>
> > fragment <- "(vsstresn < 30 or vsstresn > 130) and vstestcd = 'HR'"
> >
> > # then your program would run this code:
> >
> > df <- data.frame(vsstresn = seq(10, 200, 10), vstestcd = rep(c("HR",
> > "RH"), c(16, 4)))
> > library(sqldf)
> > sqldf(paste("select * from df where", fragment))
> vsstresn vstestcd
> 1 10 HR
> 2 20 HR
> 3 140 HR
> 4 150 HR
> 5 160 HR
Many thanks! (ok, next time I will post a data example myself.)
Is this a corollary of fortune 106: If the answer is eval(parse()),
you should reconsider the language?
--
Johannes Hüsing There is something fascinating about science.
One gets such wholesale returns of conjecture
mailto:[email protected] from such a trifling investment of fact.
http://derwisch.wikidot.com (Mark Twain, "Life on the Mississippi")
______________________________________________
[email protected] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
