r - R lubridate: How to put a lubridate timestamp in a dataframe?

Question

I am simply trying to store lubridate timestamps in a dataframe. However, they get converted in number of seconds.

Here's a reproducible example:

library(lubridate)

tsta <- ymd_hms('2021-03-16 22:59:59') # tsta for timestamp

df <- as.data.frame(matrix(ncol=1, nrow=2)) # preallocate a small dataframe
colnames(df) <- 'timestamp'

# trying to add lubridate object in it
df[1, 'timestamp'] <- tsta

So tsta looks like "2021-03-16 22:59:59 UTC", but once I put it in df it looks like

   timestamp
1 1615935599
2         NA

What's the proper way of doing this?

标签： rlubridate

Answer 1

matrix can store only a single class and POSIXct can get converted to its numeric storage values. One option is

df <- tibble(timestamp = c(tsta, ymd_hms(NA)))

-output

> df
# A tibble: 2 × 1
  timestamp          
  <dttm>             
1 2021-03-16 22:59:59
2 NA            

---

Or another way is

df <- tibble(timestamp =  ymd_hms(rep(NA, 2)))
df$timestamp[1] <- tsta
df
# A tibble: 2 × 1
  timestamp          
  <dttm>             
1 2021-03-16 22:59:59
2 NA                 

---

Or using the OP's code, the NA values are converted first

df$timestamp <- ymd_hms(df$timestamp)
df[1, 'timestamp'] <- tsta
df
            timestamp
1 2021-03-16 22:59:59
2                <NA>