CODEWITHSUNDEEP
r
tantan
tantan

Reputation: 57

How to sort by ID and then detect the difference within the same ID?

I have a dataset which contains several IDs and sample date column, like

dataframe <- data.frame(ID=c("ID1","ID2","ID3","ID4", "ID2", "ID2", "ID3","ID4", "ID5","ID1"), 
                    sample_date=c(1991-05-23, 1991-05-24,1991-05-24, 1991-05-26,1991-05-27,1991-05-28,1991-05-30,1991-05-31, 1991-06-03, 1991-06-03), 
                    sex =c(1,2,1,2,2,2,1,2,1,1), and_so_om =c(1))

and then I want to sort by the same ID and detect if the same ID has very close sample_date(e.g. within 3days).

First of all, I tried to sort the data frame by ID, and then I got the following results,

outcome <- dataframe %>% select(ID,sample_date) %>% count(ID,sample_date)

From here, I don't know how to calculate the difference of days of sample_date within the same ID.

Upvotes: 1

Views: 220

Answers (2)

ThomasIsCoding
ThomasIsCoding

Reputation: 101916

I think aggregate from base R is enough to make it

dfout <- aggregate(sample_date~ID,dataframe,function(x) min(diff(sort(x)),Inf)<=3)

such that

> dfout
   ID sample_date
1 ID1       FALSE
2 ID2        TRUE
3 ID3       FALSE
4 ID4       FALSE
5 ID5       FALSE

Upvotes: 1

Ronak Shah
Ronak Shah

Reputation: 389065

Perhaps, you can try 

library(dplyr)

n <- 3
dataframe %>%
  mutate(sample_date = as.Date(sample_date)) %>%
  arrange(ID, sample_date) %>%
  group_by(ID) %>%
  summarise(is_closest = any(diff(sample_date) <= n))

which gives

#  ID   is_close
#  <fct> <lgl>     
#1 ID1   FALSE     
#2 ID2   TRUE      
#3 ID3   FALSE     
#4 ID4   FALSE     
#5 ID5   FALSE

This checks for each ID of there is any sample_date within n days than the previous one.

data

dataframe <- data.frame(ID=c("ID1","ID2","ID3","ID4","ID2","ID2","ID3","ID4",
            "ID5","ID1"), sample_date=c("1991-05-23", "1991-05-24","1991-05-24", 
            "1991-05-26", "1991-05-27","1991-05-28","1991-05-30","1991-05-31", 
            "1991-06-03", "1991-06-03"), sex =c(1,2,1,2,2,2,1,2,1,1), and_so_om = 1)

Upvotes: 2

Related Questions