I set a custom date in excel with year, month and day because that's the level of granularity that facebook provides for their insights data. The swarm ran, but never ended and just sort of stopped running. Not sure why, but I'm looking into it.
On Wed, Sep 10, 2014 at 11:34 PM, Subutai Ahmad <[email protected]> wrote: > > I've run into this numerous times and it's frustrating. Unfortunately when > you open a CSV file in Excel it always converts to its default date format > which does not include the time. What you have to do is go into the "Format > cells" menu and type in a custom format and then save to CSV. I have > attached an example of what that would look like on a Mac. Or, just write > it out from your program in the correct format and avoid saving from Excel > altogether. > > #whymicrosoftwhy > > --Subutai > > > On Wed, Sep 10, 2014 at 6:09 PM, Scott Purdy <[email protected]> wrote: > >> I'm not sure that dates without times are supported. The error message >> lists all of the formats so just preprocess your file to match one of the >> supported date/time formats. >> >> On Wed, Sep 10, 2014 at 2:58 PM, Matthew Taylor <[email protected]> wrote: >> >>> The timestamps in the CSV do not match the time formats required: >>> >>> timestamp,likes >>> datetime,int >>> T, >>> 7/1/14,1117 >>> 7/2/14,497 >>> 7/3/14,296 >>> 7/1/14,41911 >>> 7/5/14,1129 >>> 7/6/14,301 >>> 7/7/14,2493 >>> 7/8/14,1137 >>> 7/9/14,196 >>> 7/10/14,1668 >>> >>> You'll need to update the time strings to match one of the supported >>> time formats like "2014-07-01". >>> --------- >>> Matt Taylor >>> OS Community Flag-Bearer >>> Numenta >>> >>> >>> On Wed, Sep 10, 2014 at 2:52 PM, Mika Schiller <[email protected]> >>> wrote: >>> > Yes, I've attached the raw csv file. >>> > >>> > >>> > >>> > >>> > On Wed, Sep 10, 2014 at 5:10 PM, Matthew Taylor <[email protected]> >>> wrote: >>> >> >>> >> Mika, >>> >> >>> >> Your problem might be how you're getting the CSV file out of Excel. >>> >> Can you show us the raw CSV file? I imagine that Excel is not >>> >> formatting the date outputs correctly when it exports the file to CSV. >>> >> --------- >>> >> Matt Taylor >>> >> OS Community Flag-Bearer >>> >> Numenta >>> >> >>> >> >>> >> On Tue, Sep 9, 2014 at 8:20 PM, Mika Schiller <[email protected] >>> > >>> >> wrote: >>> >> > Hey guys, >>> >> > >>> >> > I'm trying to run a Nupic prediction on some social media data. >>> Though >>> >> > when >>> >> > I try to run the swarm on the data, I get the following error: >>> >> > >>> >> > ValueError: The provided timestamp 7/1/14 is malformed. The >>> supported >>> >> > formats are: [%Y-%m-%d %H:%M:%S.%f, %Y-%m-%d %H:%M:%S:%f, %Y-%m-%d >>> >> > %H:%M:%S, >>> >> > %Y-%m-%d %H:%M, %Y-%m-%d, %m/%d/%Y %H:%M, %m/%d/%y %H:%M, >>> >> > %Y-%m-%dT%H:%M:%S.%fZ, %Y-%m-%dT%H:%M:%SZ, %Y-%m-%dT%H:%M:%S] >>> >> > >>> >> > You can take a look at the attachment to see how I have my csv file >>> >> > formatted. It's no different than the hotgym data. Excel also >>> doesn't >>> >> > seem >>> >> > to support the datetime format that the value error proposes. What >>> could >>> >> > be >>> >> > going on here? >>> >> > >>> >> > >>> >> > thnx >>> >> > >>> >> > >>> >> > Mika >>> >> >>> > >>> >>> >> >
