r dplyr group_by values collapse and paste

问题

I have a data set that looks like this

Id   Subject    Date        Vitals       Value
10   John       2001-05-29  HeartRate    65
10   John       2001-05-29  HeartRate    68
10   John       2001-05-29  BP-Arterial  48
10   John       2001-05-29  PulseRate    64
34   Pete       2005-08-15  HeartRate    68
34   Pete       2005-08-15  BP-Arterial  56
10   John       2004-09-25  HeartRate    65
10   John       2004-09-25  BP-Arterial  64
10   John       2004-09-25  PulseRate    63   
34   Pete       2007-07-21  BP-Arterial  68
34   Pete       2007-07-21  PulseRate    56

I want to do two things,

1) group by Vitals. 2) Count the number of Vitals that were measured for each ID on a specific date (ID + Date) and collapse and paste these values like this below.

Vitals      Series
HeartRate   2,1,1
BP-Arterial 1,1,1,1
PulseRate   1,1,1

The value under the Series column for HeartRate is 2, 1, 1 because HeartRate was measured

twice for ID 10 on 2001-05-29,

once for ID 34 on 2005-08-15 and

once for ID 10 on 2004-09-24

Not sure how to collapse and paste these values using dplyr any help is much appreciated.

回答1:

Count the number of Vitals that were measured for each ID on a specific date (ID + Date)

This means you need to group by all three. We can then regroup by only vitals for the final collapse:

dat %>% group_by(Vitals, Id, Date) %>%
    summarize(n = n()) %>%
    ungroup() %>%
    group_by(Vitals) %>%
    summarize(Series = paste(n, collapse = ','))
# # A tibble: 3 × 2
#        Vitals  Series
#        <fctr>   <chr>
# 1 BP-Arterial 1,1,1,1
# 2   HeartRate   2,1,1
# 3   PulseRate   1,1,1

回答2:

With dplyr and rle i.e run-length encoding see ?rle for more details

library(dplyr)

newDF = DF %>%
    group_by(Id,Date) %>%
    do(.,data.frame(Series=paste(rle(.$Vitals)$lengths,collapse=","),stringsAsFactors=FALSE)) %>%
    as.data.frame()

newDF
#  Id       Date Series
#1 10 2001-05-29  2,1,1
#2 10 2004-09-25  1,1,1
#3 34 2005-08-15    1,1
#4 34 2007-07-21    1,1

来源：https://stackoverflow.com/questions/40143046/r-dplyr-group-by-values-collapse-and-paste

标签

group-by

dplyr

collapse