CODEWITHSUNDEEP
rdata-manipulation
pssguy
pssguy

Reputation: 3505

Creating new column in R dataframe based on existing columns and grouping

I have a dataframe df of soccer team information by game (MATCHID) with these initial values

 TEAMID Venue LEAGUEPOS MATCHID
 WHU     A         5       1
 COV     H        12       1
 EVE     H        15       2
 MNU     A         2       2
 ARS     A         3       3
 LEI     H         4       3

I wish to create just one row for each game so that it would end up looking like

MATCHID HomeTeam AwayTeam HomePos AwayPos
   1       COV      WHU     12      5      etc.

so I want to create some new columns , delete others and remove duplicated rows.

I am having trouble with first stage trying

df$HomeTeam <- df$TEAMID[df$Venue == "H"]

as this produces

 TEAMID Venue LEAGUEPOS MATCHID HomeTeam
   WHU     A         5       1      COV
   COV     H        12       1      EVE
   EVE     H        15       2      LEI
   MNU     A         2       2      STH
   ARS     A         3       3      TOT
   LEI     H         4       3      WIM

With the HomeTeam just showing the sequential TEAMID for each record with a Venue = H

Upvotes: 3

Views: 3036

Answers (2)

Ramnath
Ramnath

Reputation: 55695

This can be easily achieved using the function reshape which is a part of base R.

# READ DATA
mydf = read.table(textConnection("
TEAMID Venue LEAGUEPOS MATCHID
 WHU     A         5       1
 COV     H        12       1
 EVE     H        15       2
 MNU     A         2       2
 ARS     A         3       3
 LEI     H         4       3"), 
 sep = "", header = T, colClasses = rep('character', 4))

# RESHAPE DATA
reshape(mydf, idvar = 'MATCHID', timevar = 'Venue', direction = 'wide')

Here is the output produced

  MATCHID TEAMID.A LEAGUEPOS.A TEAMID.H LEAGUEPOS.H
1       1      WHU           5      COV          12
3       2      MNU           2      EVE          15
5       3      ARS           3      LEI           4

NOTE: An alternate way to do this is to use cast and melt functions from reshape package.

require(reshape)
mydf_m = melt(mydf, id = c('MATCHID', 'Venue'))
cast(mydf_m, MATCHID ~ Venue + variable)

Upvotes: 5

Gavin Simpson
Gavin Simpson

Reputation: 174813

reshape() in base R does what you want, if a little clunkily. Here is your data:

con <- textConnection(" TEAMID Venue LEAGUEPOS MATCHID
 WHU     A         5       1
 COV     H        12       1
 EVE     H        15       2
 MNU     A         2       2
 ARS     A         3       3
 LEI     H         4       3
")
dat <- read.table(con, header = TRUE, stringsAsFactors = FALSE)
close(con)

We reshape() this, get the columns in the requested order, and update the columns names:

newdat <- reshape(dat, direction = "wide", timevar = "Venue", idvar = "MATCHID")
## reorder
newdat <- newdat[, c(1,4,2,5,3)]
names(newdat) <- c("MatchID","HomeTeam","AwayTeam","HomePos","AwayPos")

This gives us:

> newdat
  MatchID HomeTeam AwayTeam HomePos AwayPos
1       1      COV      WHU      12       5
3       2      EVE      MNU      15       2
5       3      LEI      ARS       4       3

Upvotes: 1

Related Questions