[R] Using tapply to create a new table

Kalish, Josh josh.kalish at credit-suisse.com
Fri Jan 26 19:21:18 CET 2007


Thanks for pointing out the merge function.  That gets me part of the way there.  The only thing is that I can't get the tapply() results into a format that merge() will take.  For example:

merge( set1 , tapply( set2$f1 , set2$commonField, mean ) , by="commonField" )

Gives me  "Error in names... Unused arguments..."

I'm not sure what the result of a tapply() exactly is, but it doesn't seem to be a table.

Yeah, rank amateur questions...



-----Original Message-----
From: Marc Schwartz [mailto:marc_schwartz at comcast.net] 
Sent: Friday, January 26, 2007 1:08 PM
To: Kalish, Josh
Cc: 'r-help at stat.math.ethz.ch'
Subject: Re: [R] Using tapply to create a new table

On Fri, 2007-01-26 at 12:39 -0500, Kalish, Josh wrote:
> All,
> I'm sure that this is covered somewhere, but I can't seem to find a 
> good explanation.  I have an existing table that contains information 
> grouped by date.  This is as so:
> Day		NumberOfCustomers		NumberOfComplaints
> 20060512	10040				40
> 20060513	32420				11
> ...
> I also have a table at the detail level as so:
> Day		Meal		PricePaid	UsedCupon
> 20060512	Fish		14		Y
> 20060512	Chicken	20		N
> ...
> Is there a simple way to create summaries on the detail table and then 
> join them into the first table above so that it looks like this:
> Day		NumberOfCustomers		NumberOfComplaints	AveragePricePaid
> NumberUsingCupon
> I can do a tapply to get what I want from the detail table, but I 
> can't figure out how to turn that into a table and join it back in.
> Thanks,
> Josh

Skipping the steps of using tapply() or aggregate() to get the summarized data from the second data frame, you would then use merge() to perform a SQL-like 'join' operation:

> DF.1
       Day NumberOfCustomers NumberOfComplaints
1 20060512             10040                 40
2 20060513             32420                 11

> DF.2
       Day    Meal PricePaid UsedCupon
1 20060512    Fish        14         Y
2 20060512 Chicken        20         N

> merge(DF.1, DF.2, by = "Day")
       Day NumberOfCustomers NumberOfComplaints    Meal PricePaid
1 20060512             10040                 40    Fish        14
2 20060512             10040                 40 Chicken        20
1         Y
2         N

By default, only rows matching on the 'by' argument in both data frames will be in the result. See the 'all.x' and 'all.y' arguments to handle other scenarios of including non-matching rows.

See ?merge, which BTW:


would point you to, if you are familiar with the term from relational data base operations.


Marc Schwartz

Please access the attached hyperlink for an important electr...{{dropped}}

More information about the R-help mailing list