Jacob,

Thank you so much... for the fast reply and great suggestion!  Doing this 
achieved exactly what I needed:

julia> results = query("SELECT to_char(created,'YYYY-MM-DD HH:MI:SS') from 
auditrecord order by created desc limit 10")

elapsed time: 4.2695e-5 seconds

10x1 DataFrame

|-------|-----------------------|

| Row # | to_char               |

| 1     | "2014-05-05 11:59:58" |

| 2     | "2014-05-05 11:59:57" |

| 3     | "2014-05-05 11:59:56" |

| 4     | "2014-05-05 11:59:56" |

| 5     | "2014-05-05 11:59:55" |

| 6     | "2014-05-05 11:59:55" |

| 7     | "2014-05-05 11:59:54" |

| 8     | "2014-05-05 11:59:53" |

| 9     | "2014-05-05 11:59:53" |

| 10    | "2014-05-05 11:59:52" |


On Tuesday, May 6, 2014 6:38:33 PM UTC-7, Jacob Quinn wrote:
>
> There's actually not much support in Datetime.jl for formatting/parsing 
> DateTimes (just regular Dates). (A rewrite of the package is nearing 
> completion as Dates.jl with much more solid support for formatting/parsing).
>
> In this case, I would suggest leveraging Postgres own formatting tools: 
> http://www.postgresql.org/docs/8.2/static/functions-formatting.html
>
> Cheers,
>
> -Jacob
>
>
> On Tue, May 6, 2014 at 9:35 PM, Dan B <[email protected] <javascript:>>wrote:
>
>> Im new to using Julia - thanks in advance for the help.
>>
>> The short story is that Im using ODBC and DataFrames to return query 
>> results from a Postgres database, and Julia is changing the format of my 
>> timestamps from the database, I think because it's automatically setting 
>> them as DateTime types.  Im not sure how to prevent this from happening, or 
>> change the format that it's automatically setting after the result is set.
>>
>> In the database, they are Postgres timestamp types and they look like 
>> this: *2014-04-24 10:37:10 *
>> In Julia they look like this when set in the DataFrame:  
>> *2014-04-25T09:08:10 
>> UTC *- Notice the extra T in the middle, and the timezone on the end.
>>
>> Example:
>>
>>
>> julia> results = query("SELECT created FROM auditrecord LIMIT 10")
>>
>>
>> elapsed time: 6.7873e-5 seconds
>>
>>
>> 10x1 DataFrame
>>
>> |-------|-------------------------|
>>
>> | Row # | created                 |
>>
>> | 1     | 2010-11-29T17:21:06 UTC |
>>
>> | 2     | 2010-11-29T17:22:46 UTC |
>>
>> | 3     | 2010-11-29T17:26:08 UTC |
>>
>> | 4     | 2010-11-29T17:34:55 UTC |
>>
>> | 5     | 2010-11-29T17:36:16 UTC |
>>
>> | 6     | 2010-11-29T17:36:18 UTC |
>>
>> | 7     | 2010-11-29T17:38:40 UTC |
>>
>> | 8     | 2010-11-29T17:39:40 UTC |
>>
>> | 9     | 2010-11-29T17:41:19 UTC |
>>
>> | 10    | 2010-11-29T17:43:34 UTC |
>>
>>
>> julia> eltype(results[1])
>>
>> DateTime{ISOCalendar,Zone0}
>>
>>
>> I looked at the documentation for DateTime, and I see that I can change 
>> formats, but Im not sure how to apply a format string to the array column. 
>>  (kinda wouldn't expect the following to magically change the formatting of 
>> a whole array column but it's what I want to do)  Im sure it's a newbie 
>> mistake with a simple answer.  Thanks for the help :)
>>
>> julia> f = "yyyy-MM-dd HH:mm:ss"
>>
>> "yyyy-MM-dd HH:mm:ss"
>>
>>
>> julia> datetime(f,results[1])
>>
>> ERROR: no method datetime(ASCIIString, 
>> DataArray{DateTime{ISOCalendar,Zone0},1})
>>
>
>

Reply via email to