R help - Oct 2007 - Calculating proportions from a data frame rather than a table

When one has raw data it is easy to create a table of one variable against
another and then calculate proportions
For example
a.nice.table<-table(a,b)
prop.table(a.nice.table,1)

However, I looked at several papers and created a data frame of the
aggregate data. That means I acually created a table except it is a data
frame. The first column lists the name of the first author and the year.
I cannot find how to convert the data frame to a table so I can use great
functions such as prop.table and margin.table.

Alas I tried, rowSums but it provides lousy output without listing the names
of th papers.
I have thought of going through the reshape package but I suspect that there
is an easier way to convert a data frame to a table. Is there?

-- 
Farrel Buchinsky

	[[alternative HTML version deleted]]

What am I missing here?

Cannot you just create the table from the data.frame
and apply prop.table()to it?

--- Farrel Buchinsky <fjbuch at gmail.com> wrote:
> When one has raw data it is easy to create a table
> of one variable against
> another and then calculate proportions
> For example
> a.nice.table<-table(a,b)
> prop.table(a.nice.table,1)
> 
> However, I looked at several papers and created a
> data frame of the
> aggregate data. That means I acually created a table
> except it is a data
> frame. The first column lists the name of the first
> author and the year.
> I cannot find how to convert the data frame to a
> table so I can use great
> functions such as prop.table and margin.table.
> 
> Alas I tried, rowSums but it provides lousy output
> without listing the names
> of th papers.
> I have thought of going through the reshape package
> but I suspect that there
> is an easier way to convert a data frame to a table.
> Is there?
> 
> -- 
> Farrel Buchinsky

Incidentally, the feature becomes really powerful when one uses
functions such as addmargins together with prop.table as in

coinf.table <-as.matrix(coinfection)
prop.table(coinf.table,1) # to see proportions from each HPV type per study
addmargins(coinf.table)# to  see totals within study and accross all studies
prop.table(addmargins(coinf.table,1),1)# to see proportions for all
studies added together


On 10/3/07, Farrel Buchinsky <fjbuch at gmail.com>
wrote:
> Genius! Thank you very much.
> Yes indeed I should have thought of that. For some reason I have a
> mental blank about the use of row.names and instead I repeatedly put
> that kind of data as a column in the data frame. Thank you for the
> rescue.
>

-- 
Farrel Buchinsky
GrandCentral Tel: (412) 567-7870