R help - Jul 2010 - I need help making a data.fame comprised of selected columns of an original data frame.

I must have missed something simple, but still, i don't know what.

I obtained my basic data as follows:

x <- sprintf("SELECT m_id,sale_date,YEAR(sale_date) AS
sale_year,WEEK(sale_date) AS sale_week,return_type,0.0001 +
DATEDIFF(return_date,sale_date) AS elapsed_time FROM
`merchants2`.`risk_input` WHERE DATEDIFF(return_date,sale_date) IS NOT
NULL")
moreinfo <- dbGetQuery(con, x)

I then made the data frame I want to use as follows:

fun_m_id <- function(df)
  if (length(df$elapsed_time) > 5) {
    rv = fitdist(df$elapsed_time,"exp")
    rv$mid = df$m_id[1]
    rv
  }
aaa <- lapply(split(moreinfo,list(moreinfo$m_id),drop = TRUE), fun_m_id)
m_id_default_res <- do.call(rbind, aaa)

At this point, each row in m_id_default_res corresponds to one data.frame
produced by fitdist.  When I print it, I get the output I expected.
However, I need to store only some of it into my DB.

And then, because fitdist produces a data frame that includes a lot of info
I don't need to store in the DB, I tried making a new data.frame containing
only the info I need as follows:
ndf = data.frame()
for (i in 1:length(m_id_default_res[,1])) {
  ndf$mid[i] = m_id_default_res$mid[i]
  ndf$estimate[i] = m_id_default_res$estimate[i]
  ndf$sd[i] = m_id_default_res$sd[i]
  ndf$n[i] = m_id_default_res[i]
  ndf$loglik[i] = m_id_default_res$loglik[i]
  ndf$aic[i] = m_id_default_res$aic[i]
  ndf$bic[i] = m_id_default_res$bic[i]
  ndf$chisq[i] = m_id_default_res$chisq[i]
  ndf$chisqpvalue[i] = m_id_default_res$chisqpvalue[i]
  ndf$chisqdf[i] = m_id_default_res$chisqdf[i]
}
ndf

And I get the following error:
Error in `$<-.data.frame`(`*tmp*`, "n", value =
list(0.114752782316094)) :
  replacement has 1 rows, data has 0

I need to either get rid of the columns in m_id_default_res that I don't
need, or I need to copy only those columns I need to a new data.frame.  How
do I do this.  Obviously, doing an element-wise copy, at least as I tried to
do it, doesn't work.

Thanks,

Ted

	[[alternative HTML version deleted]]

Hi,

First: it's kind of hard to play along w/o some reproducible data. To
that end, you can paste into an email the output of:

dput(moreinfo)

If there are lots of rows in `moreinfo`, just give us the first ~10-20

dput(head(moreinfo, 20))

Anyway:

<snip>
> At this point, each row in m_id_default_res corresponds to one data.frame
> produced by fitdist. ?When I print it, I get the output I expected.
> However, I need to store only some of it into my DB.
>
> And then, because fitdist produces a data frame that includes a lot of info
> I don't need to store in the DB, I tried making a new data.frame
containing
> only the info I need as follows:
> ndf = data.frame()
> for (i in 1:length(m_id_default_res[,1])) {
> ?ndf$mid[i] = m_id_default_res$mid[i]
> ?ndf$estimate[i] = m_id_default_res$estimate[i]
> ?ndf$sd[i] = m_id_default_res$sd[i]
> ?ndf$n[i] = m_id_default_res[i]
> ?ndf$loglik[i] = m_id_default_res$loglik[i]
> ?ndf$aic[i] = m_id_default_res$aic[i]
> ?ndf$bic[i] = m_id_default_res$bic[i]
> ?ndf$chisq[i] = m_id_default_res$chisq[i]
> ?ndf$chisqpvalue[i] = m_id_default_res$chisqpvalue[i]
> ?ndf$chisqdf[i] = m_id_default_res$chisqdf[i]
> }
Forget the for loop. How about:

ndf <- m_id_default[, c('mid, 'estimate', 'sd',
'loglik', 'aic',
'bic', 'chisq', 'chisqpvalue', 'chisqdf')

Having just written that, I see something strange in your for loop.
Specifically this line:
> ?ndf$n[i] = m_id_default_res[i]
m_id_default_res is a data.frame, right? Why don't you try to see what
`m_id_default_res[1]` returns.

I'm not sure that that's what your error message is coming from, but I
foresee this to be a problem anyway, if I follow your "build up" code
correctly.

Hope that helps,

-- 
Steve Lianoglou
Graduate Student: Computational Systems Biology
?| Memorial Sloan-Kettering Cancer Center
?| Weill Medical College of Cornell University
Contact Info: http://cbio.mskcc.org/~lianos/contact