Message-ID: <CACLi31XG4nnGyvjY4_ma+0wEE+oHMFjBG7mYUZZKqiS5KON9_Q@mail.gmail.com>
Date: 2011-12-12T21:32:32Z
From: Tony Stocker
Subject: Creating appropriate time axis for data
In-Reply-To: <4EE6654B.1000903@statistik.tu-dortmund.de>
2011/12/12 Uwe Ligges <ligges at statistik.tu-dortmund.de>:
> On 12.12.2011 17:44, Tony Stocker wrote:
Sorry for the double post but the first message was held for so long
that I figured there was a problem with the email address I was using
so I unsubscribed that one and resubscribed the other one.
>>
>> Hello,
>>
>> I am dealing with data stored in a database as a 'time' object. ?I
>> export the data from the database to a text file and utilize the
>> 'time_to_sec()'
>
>
> I get
>
>> time_to_sec
> Error: object 'time_to_sec' not found
>
> If it is in a package, please tell us which one you are referring to.
The time_to_sec() function is inside the database (in this case
MySQL). I'm using it solely because R doesn't want to deal with the
times as normally outputted (HH:MM:SS) because it sees them as a
character object and not a numerical or time object. By converting
these time fields to seconds when I export them from the database I
provide something that R can work with natively apparently.
>
>
>> seconds. ?However when I visualize the data in charts I do not want a
>> scale that runs from 25200-64800 seconds, but rather in the HH:MM:SS
>> format. ?Is there a relatively straight-forward and easy to use way to
>> do this?
>
>
> Yes, e.g.:
>
> t <- strptime(c("07:00:00", "18:00:00"), "%H:%M:%S")
> y <- rnorm(2)
> plot(t, x)
>
Does this method supercede the original y-axis label or supplement
it? Will the data be correctly located along the y-axis if it's in
seconds?
I thought I had looked at strptime() and it required day/date info to
be listed not just time. Based on your example I'm assuming that I
was incorrect in that assumption.
Thanks for this suggestion I'll give it a try with my data and see if
it does what I'd like it to do. I appreciate the help.
> Uwe Ligges
-Tony