R. summarizing data without merge

Question

I have a dataframe (df) of goals scored against various teams by date

gamedate teamID Gls
 1992-08-22  CHL  3
 1992-08-22  MNU  1
 1992-08-23  ARS  0
 1992-08-23  LIV  2
 1992-08-24  MNU  0
 1992-08-25  LIV  2
 1992-08-26  ARS  0
 1992-08-26  CHL  0

I wish to produce a summary table which shows the number of games played and number of games these teams have blanked the opposition on each date

gamedate   games blanks
 1992-08-22   2     0
 1992-08-23   2     1
 1992-08-24   1     1
 1992-08-25   1     0
 1992-08-26   2     2

I can get the games and blanks separately using ddply

df.a <- ddply(df,"gamedate",function(x) c(count=nrow(x)))
df.b <- ddply(subset(df,Gls==0),"gamedate",function(x) c(count=nrow(x)))

and then merger df.a and df.b to get my answer. However, I am sure there must be a more simple and elegant solution

Answer 1

You just need to use summarise:

Read the data in:

   dat <- read.table(textConnection("gamedate teamID Gls
  1992-08-22  CHL  3
  1992-08-22  MNU  1
  1992-08-23  ARS  0
  1992-08-23  LIV  2
  1992-08-24  MNU  0
  1992-08-25  LIV  2
  1992-08-26  ARS  0
  1992-08-26  CHL  0"),sep = "",header = TRUE)

and then call ddply:

ddply(dat,.(gamedate),summarise,tot = length(teamID),blanks = length(which(Gls == 0)))
    gamedate tot blanks
1 1992-08-22   2      0
2 1992-08-23   2      1
3 1992-08-24   1      1
4 1992-08-25   1      0
5 1992-08-26   2      2

Answer 2

The only thing you are missing is wrapping your functions in a data.frame() call and giving them column names... and the column names are optional :)

I'm using @joran's dat data.frame as it allowed me to test my answer.

ddply( dat, "gamedate", function(x) data.frame( 
                                      tot = nrow( x ), 
                                      blanks = nrow( subset(x, Gls == 0 ) ) 
                                              ) 
     )

BTW, my funny formatting above is just to prevent it from scrolling on the screen and to help illustrate how I'm really just bringing together the functions you already created.

Answer 3

Another solution using simple aggregate. I am using joran's dat.

agg <- aggregate(cbind(1, dat$Gls==0), list(dat$gamedate), sum)
names(agg) <- c("gamedate", "games", "blanks")
agg