Remove seconds from time

[Domingo Figueroa]

Hello all,

I am working with a large data set, I got readings every 10 seconds from the equipment. I am trying to remove the seconds from time and the letters PM/AM. I am using the following code but I am getting missing vales. Any suggestion will be greatly appreciated.

generate gassample_time = clock(var_time, "hm")

[Marcos Almeida]

The best approach is to provide data, I mean, an example, but you should try something like:

generate double gassample_time = clock(var_time,"YMDhm")

Hopfully that helps

[Nick Cox]

Discarding seconds is a rounding down operation, so one route is keep thinking of floor(). Here is some technique. Modesty aside, I am fairly experienced, but the experience boils down to this one thing: only reading the fine help can tell you what to do. Throughout, remembering that Stata's units here are milliseconds keeps you on the straight and narrow.

. clear

. set obs 1
number of observations (_N) was 0, now 1

. gen double test = clock("2019 Aug 21 20:30:40", "YMD h m s")

. format test %tc

. list

     +--------------------+
     |               test |
     |--------------------|
  1. | 21aug2019 20:30:40 |
     +--------------------+

. gen double wanted = 60000 * floor(test/60000)

. format wanted %tc

. list

     +-----------------------------------------+
     |               test               wanted |
     |-----------------------------------------|
  1. | 21aug2019 20:30:40   21aug2019 20:30:00 |
     +-----------------------------------------+

. help datetime display format

. format wanted %tcCCYY_Mon_DD_HH:MM

. list

     +----------------------------------------+
     |               test              wanted |
     |----------------------------------------|
  1. | 21aug2019 20:30:40   2019 Aug 21 20:30 |
     +----------------------------------------+

[Domingo Figueroa]

This is part of the data set I am trying to work with. I am trying to round up the time by minute in order to get the mean per minute


time ppb
11:24:05 AM 25
11:24:15 AM 44
11:24:25 AM 51
11:24:35 AM 45
11:24:45 AM 41
11:24:55 AM 37
11:25:05 AM 40
11:25:15 AM 51
11:25:25 AM 50
11:25:35 AM 44
11:25:45 AM 45
11:25:55 AM 41
11:26:05 AM 42
11:26:15 AM 39
11:26:25 AM 57
11:26:35 AM 53
11:26:45 AM 49
11:26:55 AM 49
11:27:05 AM 42

[Marcos Almeida]

Please read the FAQ, there you'll find how to share data. In short, please use - dataex - for that matter and present the data under code delimiters.

That said, assuming you have a string, you may start by dropping the last parcel.

[Nick Cox]

In addition to Marcos Almeida 's excellent advice, I add that if you want rounding up you need to use ceil() not floor(). Rounding up is not, as I understand it, how digital clocks or watches that show hours and minutes work, but that convention is just one of two possible.

Note that the syntax in #1 is not a way to get Stata to ignore what you don't want. Thus this fails to do what you want

. display clock("20:30:40", "hm")
.

because the time string doesn't match the specification. clock() is not a function to extract; it is a function to map string to numeric one-to-one.

[Marcos Almeida]

If it is a string variable, you could start with:

gen gassample_time = substr(var_time,1,8)

I'm not with my Stata now, hence I cannot test the code.

Hopefully that helps.