Insert rows in a dataframe based on a calculation

Question

I am pretty new to R and I am trying to copy a calculation done in Excel to R.
I have a data frame like this:

Component <- c("A", "B", "C")
Report_Time <- c(5781, 5781, 5781)
Interval <- c(700, 600, 800)
End_Time <- c(8281, 8281, 8281)
Start_Time <- c(800, 298, 780)
df <- data.frame(Component, Report_Time, Interval, End_Time, Start_Time)

When Printed it looks like this:

# Component Report_Time Interval    End_Time    Start_Time
#1    A         5781        700         8281        800
#2    B         5781        600         8281        298
#3    C         5781        800         8281        780

For each component, I want to populate a calculated column "Interval_Time", which is the sum of Start Time + Report_Time for First, then if it is less than End_Time the insert a row with the sum of Interval_Time (Last sum) + Interval. Repeat inserting till the sum in Interval time is less than End_Time.

# Component Report_Time Interval    End_Time    Start_Time  Interval_Time
#1   A       5781       700             8281        800         6581
#2   A       5781       700             8281        800         7281
#3   A       5781       700             8281        800         7981
#4   B       5781       1000            8281        298         6079        
#5   B       5781       1000            8281        298         7079
#6   B       5781       1000            8281        298         8079
#7   C       5781       1200            8281        780         6561
#8   C       5781       1200            8281        780         7761

I have been trying to achive this with if inside a for loop.. but haven't been succesfull.

Don't you mean "Repeat inserting till the sum in Interval time is **greater** than **End_Time**" ? "End_T2ime" looks like a typo and you want to stop when the sum is greater. — steveb, May 03 '17 at 05:59

Erdem Akkas · Accepted Answer · 2017-05-03T06:54:08.223

With data.table:

Component <- c("A", "B", "C")
Report_Time <- c(5781, 5781, 5781)
Interval <- c(700, 1000, 1200)
End_Time <- c(8281, 8281, 8281)
Start_Time <- c(800, 298, 780)
df <- data.frame(Component, Report_Time, Interval, End_Time, Start_Time)

library(data.table)
setDT(df)
df<-df[rep(1:.N,ceiling((End_Time-Start_Time-Report_Time)/Interval))]
df[,Interval_Time:=ifelse(.I==1,Start_Time+Report_Time,Start_Time+cumsum(Interval)+Report_Time-Interval),by=.(Component)]

df
Component Report_Time Interval End_Time Start_Time Interval_Time
1:         A        5781      700     8281        800          6581
2:         A        5781      700     8281        800          7281
3:         A        5781      700     8281        800          7981
4:         B        5781     1000     8281        298          6079
5:         B        5781     1000     8281        298          7079
6:         B        5781     1000     8281        298          8079
7:         C        5781     1200     8281        780          6561
8:         C        5781     1200     8281        780          7761

score 0 · Answer 2 · answered May 03 '17 at 06:28

Please check if this partial solution is useful to you. If you want to keep on adding till interval time is less than End_T2ime then you have to duplicate other rows also .

Component <- c("A", "B", "C")
Report_Time <- c(5781, 5781, 5781)
Interval <- c(700, 600, 800)
End_Time <- c(8281, 8281, 8281)
Start_Time <- c(800, 298, 780)
df <- data.frame(Component, Report_Time, Interval, End_Time, Start_Time)

df$Interval_time[1]=df[1,2]+df[1,5]
for(i in 2:nrow(df))
{

  if((df[i,2]+df[i,5]) < df[i,4])
     df$Interval_time[i]=df$Interval_time[i-1]+df[i,3]
  else
    df$Interval_time[i]=df[i,2]+df[i,5]

}

score 0 · Answer 3 · answered May 03 '17 at 07:46

Not as elegant as the one by @Erden Akkas, but since I was working on it anyway ;)

NB this method works assuming the original data frame as only one observation for each component.

df$value <- df$Start_Time + df$Report_Time

for (i in 1:nrow(df))
{
  t <- df[i,]
  val <- t$value
  repeat {
    val <- val + t$Interval
    if (val > t$End_Time) {break}
    dftmp <- df[i,]
    dftmp$value <- val
    # Insert new Record
    df <- rbind(df, dftmp) 

   }
 }
 df[with(df, order(Component)), ]

But this clearly is more procedural in nature as the ony by @Erden Akkas with data table library... But it gets the job done anyhow...

   Component Report_Time Interval End_Time Start_Time value
1          A        5781      700     8281        800  6581
4          A        5781      700     8281        800  7281
5          A        5781      700     8281        800  7981
2          B        5781      600     8281        298  6079
21         B        5781      600     8281        298  6679
22         B        5781      600     8281        298  7279
23         B        5781      600     8281        298  7879
3          C        5781      800     8281        780  6561
31         C        5781      800     8281        780  7361
32         C        5781      800     8281        780  8161

Insert rows in a dataframe based on a calculation

3 Answers3