Message-ID: <B0385C3D-EEDE-422A-BD6B-2D21C91E10E7@comcast.net>
Date: 2011-04-23T14:02:05Z
From: David Winsemius
Subject: Removing row with smallest value, for a given factor
In-Reply-To: <359072.60568.qm@web132403.mail.ird.yahoo.com>
On Apr 23, 2011, at 9:05 AM, - - wrote:
> I have a table.
> First column is a date, second column is an index and other columns
> contains some other values.
> I want to remove, for each date, the row with the smallest index (it
> is not necessarily 1).
>
> ex: in the following table, I want to remove row 1 (2013-05-12 with
> index 2) and row 8 (2013-05-13 with index 1)
>
> day index values
> 1 2013-05-12 2 xxxx
> 2 2013-05-12 3 xxxx
> 3 2013-05-12 4 xxxx
> 4 2013-05-12 5 xxxx
> 5 2013-05-12 6 xxxx
> 6 2013-05-12 7 xxxx
> 7 2013-05-12 8 xxxx
> 8 2013-05-13 1 xxxx
> 9 2013-05-13 3 xxxx
> 10 2013-05-13 4 xxxx
> 11 2013-05-13 5 xxxx
> 12 2013-05-13 6 xxxx
> 13 2013-05-13 7 xxxx
> 14 2013-05-13 8 xxxx
> 15 2013-05-13 9 xxxx
> 16 2013-05-13 10 xxxx
> 17 2013-05-13 12 xxxx
>
Consider using ave and creating a logical vector that you then negate:
> ave(dat$index, list(dat$day), FUN=function(x) x==min(x))
[1] 1 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0
dat[ -ave(dat$index, list(dat$day), FUN=function(x) x==min(x)), ]
--
David Winsemius, MD
West Hartford, CT