CODEWITHSUNDEEP
r
Mariah
Mariah

Reputation: 31

Extract rows from data frame

#Simulated data /model output
data1 <-data.frame(col1=c(10,20,30,40,50,60,70,80,90,100), 
                col2=c(2,4,6,8,10,12,14,16,20,22),
                col3=c(3,9,12,15,18,21,24,27,30,33),
                col4= c(4,8,12,16,20,24,28,32,36,40))

#Lower 95% CI from real data                    
lowdata <- as.data.frame(matrix(0,1,4))
                       lowdata[,1] <-5
                       lowdata[,2] <-34
                       lowdata[,3] <-25
                       lowdata[,4] <-30

# Higher 95% CI from real data
highdata <-as.data.frame(matrix(0,1,4))
                       highdata[,1] <- 59
                       highdata[,2] <- 60
                       highdata[,3] <- 50
                       highdata[,4] <- 49

I am new to R programming and not sure how to do the following.

I want to basically find out which ones of the data1 rows have values (for each column) that are within the Cis defined by two the two dataframes (lowdata1 and highdata1).

I have tried this, but it takes each value and not row: 

wmax <- which(data1[,1:4] < highdata1[,1:4])
wmin <- which(data1[,1:4] > lowdata1[,1:4])
w <- intersect(wmax, wmin)

How could I achieve extracting which rows in data1 basically "fit" between the CIs?

Upvotes: 2

Views: 53

Answers (1)

IceCreamToucan
IceCreamToucan

Reputation: 28675

You can check whether each element of each column is in the corresponding interval with

is.btwn <- 
  Map(function(x, low, high) x >= low & x <= high,
      data1, lowdata, highdata)

is.btwn
# $col1
#  [1]  TRUE  TRUE  TRUE  TRUE  TRUE FALSE FALSE FALSE FALSE FALSE
# 
# $col2
#  [1] FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE
# 
# $col3
#  [1] FALSE FALSE FALSE FALSE FALSE FALSE FALSE  TRUE  TRUE  TRUE
# 
# $col4
#  [1] FALSE FALSE FALSE FALSE FALSE FALSE FALSE  TRUE  TRUE  TRUE

You can use Reduce to get rows where all columns are in the interval, but in this case there are no such rows

Reduce(`&`, is.btwn)
# [1] FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE

Or the rows with any column in the corresponding interval

Reduce(`|`, is.btwn)
# [1]  TRUE  TRUE  TRUE  TRUE  TRUE FALSE FALSE  TRUE  TRUE  TRUE

All of these results can be passed to which to get indices with value TRUE

Upvotes: 1

Related Questions