Question

I am facing an issue while subsetting a data frame in R. Data frame is att2 which has a column filter_name based upon which I want to subset. The unique values for this column are below.

unique(att2[["filter_name"]])
# [1] title             Type        Operating_System         Occasion           Brand
148 Levels: Accessories Age Antennae Art_Style Aspect_ratio ... Zoom

This shows that Brand is a value for filter_name column. But when I subset the frame using below code, it gives 0 rows as below.

att3 <- subset(att2, filter_name == 'Brand')
> att3
[1] a      b         c  filter_name
<0 rows> (or 0-length row.names)

I am not able to find out the reason. Has anyone faced this kind of issue?

Answer 1

All that we can do is guess at what the source of your problem might be.

Here’s my best guess: Your “filter_name” column has whitespace in it, thus you shouldn’t actually be looking for “Brand” until you strip the whitespace.

Here’s a minimal example that reproduces your problem if my guess is correct:

First, some sample data:

mydf <- data.frame(Param =  c("   Brand   ", "Operating System", 
                              "Type ", "   Brand   ", "Type ", 
                              "Type ", "   Brand   ", "Type ", 
                              "   Brand   "), Value = 1:9)
unique(mydf[["Param"]])
# [1]    Brand         Operating System Type            
# Levels:    Brand    Operating System Type 

subset(mydf, Param == "Brand")
# [1] Param Value
# <0 rows> (or 0-length row.names)

Use print with the quote = TRUE argument to see the whitespace in your data.frame:

print(mydf, quote = TRUE)
#                Param Value
# 1      "   Brand   "   "1"
# 2 "Operating System"   "2"
# 3            "Type "   "3"
# 4      "   Brand   "   "4"
# 5            "Type "   "5"
# 6            "Type "   "6"
# 7      "   Brand   "   "7"
# 8            "Type "   "8"
# 9      "   Brand   "   "9"

If that happens to be your problem, then a quick gsub should fix it:

mydf$Param <- gsub("^\\s+|\\s+$", "", mydf$Param)
unique(mydf[["Param"]])
# [1] "Brand"            "Operating System" "Type"  

subset(mydf, Param == "Brand")
#   Param Value
# 1 Brand     1
# 4 Brand     4
# 7 Brand     7
# 9 Brand     9

You may also want to look into the strip.white argument in read.table and family which defaults to FALSE. Try re-reading in your data with strip.white = TRUE and then try your subsetting.