计算两个数据框中日期之间的差异

Question

The below is the first data frame

Account reference number    Amount  Date
A   1   1583.51 16/05/2016
B   2   4038.18 27/09/2016
C   3   1161.36 20/05/2016
C   4   732.39  24/10/2016
C   5   747.69  24/11/2016

The below is second data frame

Account reference number    Amount  Date
A   6   3062.88 03/05/2016
A   7   2619.09 03/05/2016
A   8   4743.22 09/05/2016
B   9   115.28  03/05/2016
B   10  993.14  03/05/2016
B   11  879.05  03/05/2016
C   12  50.93   03/05/2016
C   13  21.83   03/05/2016
C   14  14.55   03/05/2016

I want to find the date difference for each account by comparing two data frames. For example, if you compare the dates in data frames for account 'A' it should be -13 days as start date would be 16/05/2016 and the stop date would be 03/05/2016.

I want the date in first data frame to be checked with every date in second data frame for that account. For example, 16/05/2016 should check with 03/05/2016 and 09/05/2016.

Answer 1

Created my own sample data, since yours is hard to copy. A solution based on dplyr:

df1 = data.frame(account=c(1,2,3,4),date=seq(Sys.Date(),Sys.Date()+3,by=1),value = c(1,1,1,1))
df2 = data.frame(account=c(1,2,3,4),date=seq(Sys.Date()+2,Sys.Date()+5,by=1), value = c(2,2,2,2))

require(dplyr)

df2 = df2 %>% select(account,df2.date=date)
df1 = df1 %>% left_join(df2) %>% mutate(diff = as.numeric(date-df2.date))

INPUT

> df1
  account       date value
1       1 2017-07-17     1
2       2 2017-07-18     1
3       3 2017-07-19     1
4       4 2017-07-20     1
> df2
  account       date value
1       1 2017-07-19     2
2       2 2017-07-20     2
3       3 2017-07-21     2
4       4 2017-07-22     2

OUTPUT

> df1
  account       date value   df2.date diff
1       1 2017-07-17     1 2017-07-19   -2
2       2 2017-07-18     1 2017-07-20   -2
3       3 2017-07-19     1 2017-07-21   -2
4       4 2017-07-20     1 2017-07-22   -2

Hope this helps!

Answer 2

For simplicity I suppose that the first date frame is called a and the second b. I've created them in an abbreviated form

a <- data.frame(Account = c("A,B"), reference_number = c(1,2), Amount = c(1583.51,4038.18),  Date = c("16/05/2016","27/09/2016"))
b <- data.frame(Account = c("A,A"), reference_number = c(6,7), Amount = c(3062.88,2619.09),  Date = c("03/05/2016","03/05/2016"))

You can find differences between 2 dates in this way:

#days
difftime(strptime(b$Date[1], format = "%d/%m/%Y"),
     strptime(a$Date[1], format = "%d/%m/%Y"),units="days")

#weeks
difftime(strptime(b$Date[1], format = "%d/%m/%Y"),
     strptime(a$Date[1], format = "%d/%m/%Y"),units="weeks")

Answer 3

Using sample data based on Florian 's answer:

df1 = data.frame(account=c("A","A","B","B"),date=seq(Sys.Date(),Sys.Date()+3,by=1),value = c(1,1,1,1))
df2 = data.frame(account=c("A","A","A","B"),date=seq(Sys.Date()+2,Sys.Date()+5,by=1),value = c(2,2,2,2))

I added in several instances of each account in each data frame. This is important for obtaining correct output for your own data:

library(dplyr)
library(lubridate)
full_join(df1,df2,by="account") %>%
  mutate(diff=date.x-date.y) %>%

  account     date.x value.x     date.y value.y    diff
1       A 2017-07-17       1 2017-07-19       2 -2 days
2       A 2017-07-17       1 2017-07-20       2 -3 days
3       A 2017-07-17       1 2017-07-21       2 -4 days
4       A 2017-07-18       1 2017-07-19       2 -1 days
5       A 2017-07-18       1 2017-07-20       2 -2 days
6       A 2017-07-18       1 2017-07-21       2 -3 days
7       B 2017-07-19       1 2017-07-22       2 -3 days
8       B 2017-07-20       1 2017-07-22       2 -2 days 

Answer 4

You can use plyr and dplyr package to get your desired output. It first sorts the combined data frame, then calculates time difference between first date and date in for each row within each group. After that it finds the maximum for each group and at the end gets rid of added column.

df <- rbind(df1,df2)
df$Date <- as.Date(df$Date, "%d/%m/%Y")

library(dplyr)

df <- df %>% 
         arrange(Account, Date)

library(plyr)


plyr::ddply((df), .(Account), transform, 
      Date_1 = Date[1],
      change = abs((Date - Date[1]))) %>% 
             dplyr::group_by(Account) %>% 
             dplyr::slice(which.max(change)) %>%
             dplyr::select(-Date_1)


# Source: local data frame [3 x 5] 
# Groups: Account [3] 
#  
# # A tibble: 3 x 5 
#   Account reference.number  Amount       Date   change 
#    <fctr>            <int>   <dbl>     <date>   <time> 
# 1       A                1 1583.51 2016-05-16  13 days 
# 2       B                2 4038.18 2016-09-27 147 days 
# 3       C                5  747.69 2016-11-24 205 days

Data

df1 <- structure(list(Account = structure(c(1L, 2L, 3L, 3L, 3L), .Label = c("A", 
"B", "C"), class = "factor"), reference.number = 1:5, Amount = c(1583.51, 
4038.18, 1161.36, 732.39, 747.69), Date = structure(c(1L, 5L, 
2L, 3L, 4L), .Label = c("16/05/2016", "20/05/2016", "24/10/2016", 
"24/11/2016", "27/09/2016"), class = "factor")), .Names = c("Account", 
"reference.number", "Amount", "Date"), class = "data.frame", row.names = c(NA,-5L))

df2 <- structure(list(Account = structure(c(1L, 1L, 1L, 2L, 2L, 2L, 
3L, 3L, 3L), .Label = c("A", "B", "C"), class = "factor"), reference.number = 6:14, 
Amount = c(3062.88, 2619.09, 4743.22, 115.28, 993.14, 879.05, 
50.93, 21.83, 14.55), Date = structure(c(1L, 1L, 2L, 1L, 
1L, 1L, 1L, 1L, 1L), .Label = c("03/05/2016", "09/05/2016"
), class = "factor")), .Names = c("Account", "reference.number", 
"Amount", "Date"), class = "data.frame", row.names = c(NA, -9L))