如何将数据框的奇数列转换为POSIXct格式？

Let’s break down why your current approaches are failing and fix this step by step:

• Running as.POSIXct(timestamptest2[, odd_indexes]) directly throws an error because as.POSIXct.default doesn’t know how to process a data frame input—it expects a single vector, not multiple columns at once.
• Using apply() converts your data frame columns into a matrix, which often coerces factor columns to their underlying integer codes instead of keeping the date string values. That’s why you get the "character string is not in a standard unambiguous format" error—you’re accidentally trying to convert integers, not actual date strings.

Here are two reliable, tested solutions:

1. Base R Approach with lapply() #

lapply() processes each column individually without coercing to a matrix, preserving the original column structure. We’ll also handle factor columns (a common hidden issue here) by converting them to characters first:

# Define your odd column indexes (example: c(1,3,5) for a 5-column data frame)
odd_indexes <- seq(1, ncol(timestamptest2), by = 2)

# Convert target columns to POSIXct
timestamptest2[, odd_indexes] <- lapply(timestamptest2[, odd_indexes], function(col) {
  # Convert factors to character strings if needed
  if (is.factor(col)) {
    col <- as.character(col)
  }
  # Specify timezone to match your expected output (CEST maps to "Europe/Berlin")
  as.POSIXct(col, tz = "Europe/Berlin")
})

2. Tidyverse Approach with dplyr::mutate_at() #

If you use the tidyverse ecosystem, this syntax is cleaner and more readable:

library(dplyr)

odd_indexes <- seq(1, ncol(timestamptest2), by = 2)

timestamptest2 <- timestamptest2 %>%
  mutate_at(vars(odd_indexes), ~as.POSIXct(., tz = "Europe/Berlin"))

Bonus: Handle Non-Standard Date Formats #

If your date strings don’t follow R’s default expected format, add the format parameter to as.POSIXct(). For example, if your dates look like 01-05-2018 15:00:16, use:

as.POSIXct(col, tz = "Europe/Berlin", format = "%d-%m-%Y %H:%M:%S")

Verify the Fix #

Check the column types after conversion to confirm success:

str(timestamptest2[, odd_indexes])

You should see each column listed as POSIXct with the correct timezone attached.

内容的提问来源于stack exchange，提问作者tfkLSTM