square brackets multiple columns R

Question

I am flummoxed. I am trying to isolate certain rows of df according to values in two columns. As always i try this in practice data first. My code works fine.

data1<-df2[df2$fruit=="kiwi" |  df2$fruit=="orange" | df2$fruit=="apple"  & (df2$dates>= "2010-04-01" & df2$dates<  "2010-10-01"), ]

when I try the same code on my real data, it doesn't work. It collects the "fruits" I need, but ignores my date range request.

 data1<-lti_first[lti_first$hai_atc=="C10AA01" | lti_first$hai_atc=="C10AA03" | lti_first$hai_atc=="C10AA04" | lti_first$hai_atc=="C10AA05" | lti_first$hai_atc=="C10AA07" | lti_first$hai_atc=="C10AB02" |lti_first$hai_atc=="C10AA04" |lti_first$hai_atc=="C10AB08" | lti_first$hai_atc=="C10AX09" & (lti_first$date_of_claim >= "2010-04-01" & lti_first$date_of_claim<"2010-10-01"), ]

the structure of the variables in my practice data and real data are the exact same. Fruits/hai_atc are factors in both dfs, dates are as.Dates in both dfs.

in an effort to get around this I've tried subsetting my data instead, but that won't work for me either (but does work on practice data)

x<-subset(lti_first, hai_atc=="V07AY03" | hai_atc=="A11JC94" & (date_of_claim>="2010-04-01" & date_of_claim<"2010-10-01"))

What am I doing wrong? To me, my code looks identical!

sample df

names<-c("tom", "mary", "tom", "john", "mary",
 "tom", "john", "mary", "john", "mary", "tom", "mary", "john", "john")
dates<-as.Date(c("2010-02-01", "2010-05-01", "2010-03-01", 
"2010-07-01", "2010-07-01", "2010-06-01", "2010-09-01",
 "2010-07-01", "2010-11-01", "2010-09-01", "2010-08-01", 
"2010-11-01", "2010-12-01", "2011-01-01"))
fruit<-as.character(c("apple", "orange", "banana", "kiwi",
 "apple", "apple", "apple", "orange", "banana", "apple",
 "kiwi", "apple", "orange", "apple"))
age<-as.numeric(c(60,55,60,57,55,60,57,55,57,55,60,55, 57,57))
sex<-as.character(c("m","f","m","m","f","m","m",
 "f","m","f","m","f","m", "m"))
df2<-data.frame(names,dates, age, sex, fruit)
df2


dput(df2)
structure(list(names = structure(c(3L, 2L, 3L, 1L, 2L, 3L, 1L, 
2L, 1L, 2L, 3L, 2L, 1L, 1L), .Label = c("john", "mary", "tom"
), class = "factor"), dates = structure(c(14641, 14730, 14669, 
14791, 14791, 14761, 14853, 14791, 14914, 14853, 14822, 14914, 
14944, 14975), class = "Date"), age = c(60, 55, 60, 57, 55, 60, 
57, 55, 57, 55, 60, 55, 57, 57), sex = structure(c(2L, 1L, 2L, 
2L, 1L, 2L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 2L), .Label = c("f", 
"m"), class = "factor"), fruit = structure(c(1L, 4L, 2L, 3L, 
1L, 1L, 1L, 4L, 2L, 1L, 3L, 1L, 4L, 1L), .Label = c("apple", 
"banana", "kiwi", "orange"), class = "factor")), .Names = c("names", 
"dates", "age", "sex", "fruit"), row.names = c(NA, -14L), class = "data.frame")

**real data too big to put in dput, here's an str instead

str(sample_lti_first)
'data.frame':   20 obs. of  5 variables:
 $ hai_dispense_number: Factor w/ 53485 levels "Patient HAI0000017",..: 22260 22260 2527 24311 24311 24311 24311 13674 13674 13674 ...
 $ sex                : Factor w/ 4 levels "F","M","U","X": 2 2 2 1 1 1 1 1 1 1 ...
 $ hai_age            : int  18 18 27 40 40 40 40 28 28 28 ...
 $ date_of_claim      : Date, format: "2009-10-09" "2009-10-09" "2009-10-18" ...
 $ hai_atc            : Factor w/ 1038 levels "","A01AA01","A01AB03",..: 144 76 859 80 1009 1009 859 81 1008 859 ...

Answer 1

Does this work?

data1 <- subset(lti_first,
  (hai_atc %in% c("C10AA01", "C10AA03", "C10AA04", "C10AA05", "C10AA07",
                  "C10AB02", "C10AA04", "C10AB08", "C10AX09")) & 
  (date_of_claim >= as.Date("2010-04-01") & date_of_claim < as.Date("2010-10-01")))

Note the use of %in% and as.Date.

Answer 2

I think it's important to expand on @Aaron's comment. The problem you ran into was caused by a lack of parentheses around all the OR statements that using %in% avoided, and not that OR statements don't work in the extract function [. Your toy example actually didn't work exactly how you wanted - there was an orange fruit with the date 2010-12-01. That other problems didn't crop up was just chance.

The way to read the boolean logic in this code

df2[df2$fruit=="kiwi" |  df2$fruit=="orange" | df2$fruit=="apple"  & (df2$dates>= "2010-04-01" & df2$dates<  "2010-10-01"), ]

is:

I want all rows of df2 where fruit is kiwi, all rows where fruit is orange, and all rows where fruit is apple and the dates are between 3/31/2010 and 10/1/2010.

That is what you got - only the apples were truncated to the appropriate date range. There were actually no kiwis outside the date range in the toy dataset.

Now add a pair of parentheses:

df2[(df2$fruit=="kiwi" |  df2$fruit=="orange" | df2$fruit=="apple")  & (df2$dates >= "2010-04-01" & df2$dates <  "2010-10-01"), ]

This code says:

I want all the rows of df2 where fruit is either kiwi, orange, or apple and the dates are between 3/31/2010 and 10/1/2010.

That all being said, %in% is definitely the way to go.