Sum a filtered column

Question

I am trying to sum up a filtered column to count the number of occurrences. How can i sum the character data type.

bowler           dismissal_type
TS Mills         bowled
TS Mills         bowled
A Nehra          bowled
Pj cummins       bowled
TS Mills         bowled
A Nehra          bowled

I tried to sum the dismissal_type but throws me an errors saying can't sum character type

bowled <- innings%>%
 filter(dismissal_type == "bowled")%>%
 group_by(bowler)%>%
 summarize_each(bowled = sum(dismissal_type == "bowled"))%>%
 top_n()

It seems simple but can't get it done. All i want to output is:

bowler           dismissal_type  n 
TS Mills         bowled          3
A Nehra          bowled          2
Pj cummins       bowled          1

Maybe `innings %>% filter(dismissal_type == "bowled") %>% count(bowler)`? — Jaap, May 25 '19 at 10:47

niko · Answer 1 · 2019-05-25T10:59:26.587

0

You could do

# Assuming the data frame is called bowler

# Define the subset
my_subset <- bowler$dismissal_type == "bowled"
# Count the occurrences of column bowler 
my_counts <- as.integer(ave(bowler$bowler, bowler$bowler, FUN = length))
# [1] 3 3 2 1 3 2
# Initialize the column n (needed in case the subset is strict)
bowler$n <- 1L  # or any other number, na_integer, etc.
bowler$n[my_subset] <- my_counts[my_subset]
#       bowler dismissal_type n
# 1   TS Mills         bowled 3
# 2   TS Mills         bowled 3
# 3    A Nehra         bowled 2
# 4 Pj cummins         bowled 1
# 5   TS Mills         bowled 3
# 6    A Nehra         bowled 2

edited May 25 '19 at 10:59

answered May 25 '19 at 10:48

niko

5,253
1
12
32

I would use `tapply` instead of `ave`. – Rui Barradas May 25 '19 at 11:33
@RuiBarradas But you'd need another step to match the lengths right? (I'm not really familiar with `tapply`) – niko May 27 '19 at 15:00
No because in the expected output the OP wants only one row per count, not the counts in all input rows. All that is needed is a summary table. – Rui Barradas May 27 '19 at 15:14

Sum a filtered column

1 Answers1