Mismatch between date formatted as string and numeric date

[Mario Liebensteiner]

Dear all,

I am processing hourly time series data (DD.MM.YYYY hh:mm), where the date variable is formatted as a string and looks the following:

. list datestring

datestring

1. 01.04.2018 00:00
2. 01.04.2018 01:00
3. 01.04.2018 02:00
4. 01.04.2018 03:00
5. 01.04.2018 04:00

To convert the string variable into a date variable, I run the following code:

gen Datetime = clock(datestring, "DMYhm")
format Datetime %tc


This is what I get:

. list Datetime datestring

+---------------------------------------+
| Datetime datestring |
|---------------------------------------|
1. | 01apr2018 00:00:19 01.04.2018 00:00 |
2. | 01apr2018 00:59:18 01.04.2018 01:00 |
3. | 01apr2018 02:00:28 01.04.2018 02:00 |
4. | 01apr2018 02:59:27 01.04.2018 03:00 |
5. | 01apr2018 04:00:37 01.04.2018 04:00 |
+---------------------------------------+


However, "Datetime" and "datestring" are not equivalent. For example, in row 2 the clock time is 00:59:18, whereas it should be 00:01:00.

I believe that this "mismatch" stems from the missing seconds in the variable "datestring", but I do not know how to fix this. Any help would be much appreciated!

Many thanks in advance!

Mario

[Andrew Musau]

See

help precision

For time variables, you need

gen double Datetime = clock(datestring, "DMYhm")

[Nick Cox]

help datetime

gives this advice. It is a formidable help file, but most of what you need to know about dates is in there somewhere. or a link is given. The main exceptions are to do with weeks, which seemingly you don't have to worry about.

[Mario Liebensteiner]

Dear Nick and Andrew,
Thank you indeed!
Best,
Mario