# [R] computation on a table

Claus Atzenbeck claus.atzenbeck at freenet.de
Sat Nov 12 23:41:33 CET 2005

```Hello,

I have a table (1) of the form

q1 q3 q4 q8 q9
A  5  2  0  1  3
B  2  0  2  4  4

I have another table (2):

q1 q2 q3 q4 q5 q6 q7 q8 q9
C 10  7  4  2  6  9  3  1  2

I would like to divide the numbers in table (1) by the number of the
appropriate column in table (2):

q1   q3   q4   q8   q9
A  5/10  2/4  0/2  1/1  3/2
B  2/10  0/4  2/2  4/1  4/2

The result would look lie this:

q1   q3   q4   q8   q9
A   0.5  0.5    0    1  1.5
B   0.2    0    1    4    2

BACKGROUND: I have a data frame with measured times for answering
questions. I want to know how many PERCENT of the answers are wrong,
caused by reason A or B.

This gives me the subset of false answers. The table looks like table (1):

fail <- subset(questions, type=="wrong")
fail\$qid <- factor(fail\$qid)
failtab <- table(fail\$failtype, fail\$qid)

The following gives me information about how often a specific question
was asked. This is similar to table (2) above.

count <- table(questions\$failtype, questions\$qid)
count <- colSums(count)

One solution would be to delete the line that calls factor(...) on the
subset and calculate failtab/count. However, then I have the problem
that I have to get rid of all columns of the table that have '0' in all
rows.

Thanks for any hint.
Claus

```