Conditional Sum in R | 易学教程

问题

I have a data frame that is 200 rows by 6 columns. I am interested in computing the total times that a value in Col A is less than a specific number. The number can be hard coded. I do not know where to begin...

回答1:

For a slightly more complex problem, use the "which" to tell the "sum" where to sum: if DF is the data frame:

      Ozone Solar.R  Wind Temp Month Day
 1      41     190  7.4   67     5   1
 2      97     267  6.3   92     7   8   
 3      97     272  5.7   92     7   9

Example: sum the values of Solar.R (Column 2) where Column1 or Ozone>30 AND Column 4 or Temp>90

sum(DF[which(DF[,1]>30 & DF[,4]>90),2])

回答2:

To count how many values are below some number you could use ?sum

sum( df$columnA < NUMBER )

回答3:

Just using sum on your condition will work. Logical values get converted to 0 for FALSE and 1 for TRUE so summing over a logical tells you how many values are TRUE.

 dat <- as.data.frame(matrix(1:36,6,6))
 colnames(dat) <- paste0("Col", LETTERS[1:6])
 dat$ColA
# [1] 1 2 3 4 5 6
 dat$ColA < 3
# [1]  TRUE  TRUE FALSE FALSE FALSE FALSE
 sum(dat$ColA < 3)
# [1] 2

回答4:

While the answer sum( df$columnA < NUMBER ) is correct it might be better to expand on it a little.

Say if you'd like to sum the values instead of counting you could use:

sum(df[df$columnA < Number,]$columnA)

Or if there is NA values use:

sum(df[df$columnA < Number,]$columnA, na.rm=TRUE)
sum(df[(df$columnA < Number)&(!is.na(df$columnA)),]$columnA)

Basically what happens there is that you create a boolean vector of columnA which has TRUE/FALSE based on your conditional. Then you're taking a subset of the original dataframe and using it in this case to do summation of columnA.

Here's an example you can use to try it out:

df = data.frame(colA=c(1, 2, 3, 4, NA), colB=c('a', NA, 'c', 'd', 'e'))

# Count
sum(df$colA) # NA
sum(df$colA, na.rm=TRUE) # 10 This is actually sum of values since colA wasn't turned into vector of booleans
sum(df$colA > 0, na.rm=TRUE) # 4
sum(df$colA > 2, na.rm=TRUE) # 2
sum((df$colA > 2) & (df$colB == 'd'), na.rm=TRUE) # 1

# Sum of values
sum(df$colA, na.rm=TRUE) # 10
sum(df[df$colA > 0,]$colA, na.rm=TRUE) # 10
sum(df[df$colA > 2,]$colA, na.rm=TRUE) # 7
bn_vector = (df$colA > 2)&(df$colB=='d') # Boolean vector
sub_df = df[bn_vector,] # Subset of the dataframe. Leaving the second argument in [] empty uses all the columns
sub_df_colA = df[bn_vector, 'colA'] # Content of column 'colA' which is vector of numbers
sum(sub_df$colA) # 4
sum(sub_df_colA) # 4

回答5:

Ozone<-c(41,97,97)

Solar.R<-c(190,267,272)

Wind<-c(7.4,6.3,5.7)

Temp<-c(67,92,92)

Month<-c(5,7,7)

Day<-c(1,8,9)

tbl<-data.frame(Ozone,Solar, Wind , Temp,Month, Day)

tbl

Ozone | Solar.R | Wind | Temp | Month | Day 1 41 | 190 | 7.4 | 67 | 5 | 1 2 97 | 267 | 6.3 | 92 | 7 | 8 3 97 | 272 | 5.7 | 92 | 7 | 9

sum(tbl$Temp) / sum(!is.na(tbl$Temp))

[1] 84

来源：https://stackoverflow.com/questions/10827705/conditional-sum-in-r

标签

conditional

sum