Display specific columns in r based on value

Question

There are a few questions concerning this, but my problem doesn't seem to fit in with the others: I have a column called "Time" that can be either 0 or 1. Based on this value, I'd use the values in another column on the same row.

For instance, I want to display a summary of all words in column Node but only for those columns whose Time == 1. In other words: if Time==1, display value of column Node on the same row.

I tried:

summary(d[d$Time=="1", ]$Node)

However, this also displays words that do not belong to Time==1. These words do have a calue of 0 then.

Example:

Let's say this is d:

Time - Node
 0      Banana
 1      Apple
 1      Banana
 0      Strawberry
 1      Apple
 0      Coconut
 1      Peach

The output should be:

Apple 2
Banana 1
Peach 1

And not

Apple 2
Banana 1
Peach 1
Strawberry 0
Coconut 0

You could also aggreagte and subset: `x <- aggregate(Time ~ Node, df, FUN = sum); x[x$Time > 0,]` — talat, Dec 24 '14 at 11:34
Or possibly a more canonical dupe [**here**](http://stackoverflow.com/questions/1195826/dropping-factor-levels-in-a-subsetted-data-frame-in-r) with several ways to drop unused levels. — Henrik, Dec 24 '14 at 11:36
Or just for fun `library(data.table) ; setDT(d)[, sum(Time)[sum(Time) > 0], keyby = Node]` — David Arenburg, Dec 24 '14 at 11:36
@DavidArenburg, I guess it would if they were factors. Meh. Who stores character variables as factors these days.... — A5C1D2H2I1M1N2O1R2T1, Dec 24 '14 at 11:43
(`soread` has `stringsAsFactors` set to `FALSE` -- should remember that in the future (for a function I wrote...)) — A5C1D2H2I1M1N2O1R2T1, Dec 24 '14 at 11:44
@DavidArenburg, I _think_ your data.table code could be optimized :D i.e. subset first and only calculate the sum for Time > 0 so that it would be: `setDT(df)[Time > 0, sum(Time), keyby = Node]` — talat, Dec 24 '14 at 11:47
@docendodiscimus, yeah are right. I guess I can say the same to you: `aggregate(Time ~ Node, df[df$Time > 0, ], FUN = sum)` — David Arenburg, Dec 24 '14 at 11:51

score 0 · Accepted Answer · answered Dec 24 '14 at 11:31

0

You could use droplevels to remove the levels of the factor variable Node while subsetting

summary(droplevels(d[d$Time=="1", ]$Node))
#  Apple Banana  Peach 
#  2      1      1

Or

 summary(factor(d[d$Time=='1', ]$Node))

answered Dec 24 '14 at 11:31

akrun

874,273
37
540
662

Display specific columns in r based on value

1 Answers1