Reputation: 3230
I have two columns in data frame
2010 1
2010 1
2010 2
2010 2
2010 3
2011 1
2011 2
I want to count frequency of both columns and get the result in this format
y m Freq
2010 1 2
2010 2 2
2010 3 1
2011 1 1
2011 2 1
Upvotes: 35
Views: 67128
Reputation: 1732
Here another approach that I found here:
df<- structure(list(y = c(2010L, 2010L, 2010L, 2010L, 2010L, 2011L,
2011L), m = c(1L, 1L, 2L, 2L, 3L, 1L, 2L)), .Names = c("y", "m"
), class = "data.frame", row.names = c(NA, -7L))
Two options:
aggregate(cbind(count = y) ~ m,
data = df,
FUN = function(x){NROW(x)})
or
aggregate(cbind(count = y) ~ m,
data = df,
FUN = length)
Upvotes: -1
Reputation: 6778
Here is a simple base R
solution using table()
and as.data.frame()
df2 <- as.data.frame(table(df1))
# df2
y m Freq
1 2010 1 2
2 2011 1 1
3 2010 2 2
4 2011 2 1
5 2010 3 1
6 2011 3 0
df2[df2$Freq != 0, ]
# output
y m Freq
1 2010 1 2
2 2011 1 1
3 2010 2 2
4 2011 2 1
5 2010 3 1
Data
df1 <- structure(list(y = c(2010L, 2010L, 2010L, 2010L, 2010L, 2011L,
2011L), m = c(1L, 1L, 2L, 2L, 3L, 1L, 2L)), .Names = c("y", "m"
), class = "data.frame", row.names = c(NA, -7L))
Upvotes: 4
Reputation: 99391
I haven't seen a dplyr answer yet. The code is rather simple.
library(dplyr)
rename(count(df, y, m), Freq = n)
# Source: local data frame [5 x 3]
# Groups: V1 [?]
#
# y m Freq
# (int) (int) (int)
# 1 2010 1 2
# 2 2010 2 2
# 3 2010 3 1
# 4 2011 1 1
# 5 2011 2 1
Data:
df <- structure(list(y = c(2010L, 2010L, 2010L, 2010L, 2010L, 2011L,
2011L), m = c(1L, 1L, 2L, 2L, 3L, 1L, 2L)), .Names = c("y", "m"
), class = "data.frame", row.names = c(NA, -7L))
Upvotes: 16
Reputation: 1294
A more idiomatic data.table version of @ugh's answer would be:
library(data.table) # load package
df <- data.frame(y = c(rep(2010, 5), rep(2011,2)), m = c(1,1,2,2,3,1,2)) # setup data
dt <- data.table(df) # transpose to data.table
dt[, list(Freq =.N), by=list(y,m)] # use list to name var directly
Upvotes: 8
Reputation: 8634
Using sqldf
:
sqldf("SELECT y, m, COUNT(*) as Freq
FROM table1
GROUP BY y, m")
Upvotes: 5
Reputation: 789
library(data.table)
oldformat <- data.table(oldformat) ## your orignal data frame
newformat <- oldformat[,list(Freq=length(m)), by=list(y,m)]
Upvotes: 3
Reputation: 42343
If you had a very big data frame with many columns or didn't know the column names in advance, something like this might be useful:
library(reshape2)
df_counts <- melt(table(df))
names(df_counts) <- names(df)
colnames(df_counts)[ncol(df_counts)] <- "count"
df_counts
y m count
1 2010 1 2
2 2011 1 1
3 2010 2 2
4 2011 2 1
5 2010 3 1
6 2011 3 0
Upvotes: 4
Reputation: 4643
If your data is dataframe df
with columns y
and m
library(plyr)
counts <- ddply(df, .(df$y, df$m), nrow)
names(counts) <- c("y", "m", "Freq")
Upvotes: 42