Hi Chris,
Would the data have any value to you if it wasn't timestamped?
Is there any way of getting the file created with a self describing format like JSON or XML, so you know by looking at the file what each value represents?
I'd be tempted to look at a NoSQL database since these have a more flexible schema and are a lot less bothered about some records having different structure to others.
Cheers,
Ewan
On 29 September 2014 14:25, Chris Green cl@isbd.net wrote:
I have a Beaglebone Black which gathers various bits of data on our boat in France. At present these comprise some temperatures and battery voltages and currents but there might be other things added.
I want to store these in a sqlite database on my desktop machine here at home. The mechanics of doing this are easy enough, I store the data in a file named for the date and time on the Beaglebone Black and at intervals my desktop machine mounts the directory where these files are saved (using sshfs), reads the data, inserts it into the database and then deletes the file. This means that if the connection fails for a while then nothing is lost, it's probably not by any means perfect but it's good enough, my life doesn't depend on the data! :-)
At the moment the receiving end just assumes there are ten values sent and that they are in a specific order. E.g. there's a file there at the moment whose name is "2014-09-29_15:05.od" and whose contents are:-
'2014-09-29 15:05', 22.125, 22.75, 13.6285714286, 13.8876923077, 34.36875, 0.253125, 14.85, 8.803125, 8I'm not really happy with this because it means that any changes I do have to be 'synchronised' at the two ends to keep their knowledge of which value is which in step.
Is there a standard way of labelling values in this sort of situation? Should I send a file something like:-
DateTime, '2014-09-29 15:05' InsideTemperature, 22.125 OutsideTemperature, 22.75 LeisureVolts, 13.6285714286 StarterVolts, 13.8876923077 LeisureAmps1, 34.36875 etc.Then the first column of this data is the database field name and the second column is the data. If a field name doesn't exist then the receiving program/script will create a new column. This rather depends on there always being a DateTime field so maybe the receiving script could reject files with no DateTime value.
Does this seem a reasonable way to do it or are there better ways, or should I just stay with the existing set-up where the order is significant?
-- Chris Green
main@lists.alug.org.uk http://www.alug.org.uk/ http://lists.alug.org.uk/mailman/listinfo/main Unsubscribe? See message headers or the web site above!