r - "Error in as.Date.numeric(e) : 'origin' must be supplied" when adding rows

Question

I have a dataframe, df, like so:

df
        date periodNumber  value
1 2020-02-02            0   814
2 2020-02-09            0  1592
3 2020-02-09            1  1808
4 2020-02-16            0  2746
5 2020-02-16            1  2646
6 2020-02-16            2  2993

str(df)
'data.frame':   6 obs. of  3 variables:
 $ date        : Date, format: "2020-02-02" "2020-02-09" "2020-02-09" "2020-02-16" ...
 $ periodNumber: int  0 0 1 0 1 2
 $ value       : num  814 1592 1808 2746 2646 2993

As can ben seen, df$date is in the Date format. I have a for loop to add rows to this dataframe, like so:

for(i in 1:nrow(df)){
  df[nrow(df)+1,] <- c(df$date[i]+1, df$periodNumber[i], df$periodNumber[i])
}

However, this gives the error Error in as.Date.numeric(e) : 'origin' must be supplied. Why is this? The loop works if I remove the Date column. I've also tried versions of adding as.Date() to the for loop, but that gives the same error. For example, the code below gives the same error:

for(i in 1:nrow(df)){
  df[nrow(df)+1,] <- c(as.Date(df$date[i], origin='1970-01-01')+1, df$periodNumber[i], df$periodNumber[i])
}

question from:https://stackoverflow.com/questions/65850043/error-in-as-date-numerice-origin-must-be-supplied-when-adding-rows

Answer 1

The c in the for loop wants to coerce the three values into a single format.

In ?c we can read:

The output type is determined from the highest type of the components in the hierarchy NULL < raw < logical < integer < double < complex < character < list < expression. [...] factors are treated only via their internal integer codes.

I'm actually not sure where "Date" where date can be categorized here. One might expect that it would be coerced to numeric, but it's date. And because there is no origin= supplied (where should it actually be defined??) it throws an error.

This is actually interesting, since

c(1, as.Date("2020-01-01"))
# [1]     1 18262

but

c(as.Date("2020-01-01"), 1)
# Error in as.Date.numeric(e) : 'origin' must be supplied

This is at least an inconsistency, maybe a bug, I could not find any documentation.

Anyway, use data.frame to allow for multiple classes.

for(i in 1:nrow(df)){
  df[nrow(df)+1,] <- data.frame(df$date[i]+1, df$periodNumber[i], df$periodNumber[i])
}
#          date periodNumber value
# 1  2020-02-02            0   814
# 2  2020-02-09            0  1592
# 3  2020-02-09            1  1808
# 4  2020-02-16            0  2746
# 5  2020-02-16            1  2646
# 6  2020-02-16            2  2993
# 7  2020-02-03            0     0
# 8  2020-02-10            0     0
# 9  2020-02-10            1     1
# 10 2020-02-17            0     0
# 11 2020-02-17            1     1
# 12 2020-02-17            2     2

---

Data:

df <- structure(list(date = structure(c(18294, 18301, 18301, 18308, 
18308, 18308), class = "Date"), periodNumber = c(0L, 0L, 1L, 
0L, 1L, 2L), value = c(814L, 1592L, 1808L, 2746L, 2646L, 2993L
)), row.names = c("1", "2", "3", "4", "5", "6"), class = "data.frame")