[postgis-users] Length of Bigint, Double and Text fields

Gerald Creager gerry.creager at tamu.edu
Tue Dec 1 09:30:27 PST 2009


My experience in the past (including some .dbf imports by several 
systems, including legaccy dBase IV, was that in fact, these are dB-IV 
.dbf's.  The difference is small, as noted below, between v. III and v. 
IV.

gerry

Peter Hopfgartner wrote:
> Chris Hermansen wrote:
>> Peter;
>>
>> The canonical description of DBase files seems to be
>>
>> http://www.clicketyclick.dk/databases/xbase/format/
>>
>>   
> Well, first of all, I do not know which dBase version is targeted by 
> ArcGIS.
> 
> Second, I doubt the exact values in 
> http://www.clicketyclick.dk/databases/xbase/format/data_types.html#DATA_TYPES. 
> 
> 
>  From our tests, both dbase III and dBase IV can handle text fields <= 
> 254 characters (and not < 254).
> Further, dBase III can handle numerics with <= 19 digits and not < 18 
> digits. dBase IV limits this to <= 20 digits.
> 
> The above mentioned tool identifies ArcGIS generated files as dBase IV, 
> in clear conflict on what is said in the above mentioned Wikipedia article.
> 
> Cheers,
> 
> Peter
>> On Tue, 2009-12-01 at 17:28 +0100, Peter Hopfgartner wrote:
>>  
>>> Mark Cave-Ayland wrote:
>>>    
>>>> Peter Hopfgartner wrote:
>>>>
>>>>      
>>>>> Dear PostGIS devels,
>>>>>
>>>>> since we right in the middle of comparing interoperability of shape 
>>>>> files between some FOSS programs and ArcGIS, we encoutered also the 
>>>>> following:
>>>>>
>>>>> gvSIG chokes on the Bigint as produced by pgsql2shp. Looking at 
>>>>> ArcGIs generated files, the max length for numbers is 19, compared 
>>>>> to 20 for integers and 32 for floating points as in pgsql2shp. This 
>>>>> seems to be confirmed by some Borland tools, when configured for 
>>>>> dBase III+.
>>>>>
>>>>> Another limit seems to be that of text fields, that must be no 
>>>>> longer then 254. This is the maximum as given by ArcGIS and by the 
>>>>> Borland tool.
>>>>>
>>>>> Another point is Boolean: ArcGIS seems to not allow the creation of 
>>>>> boolean fields and the Borland Tool ("Database Desktop") says that 
>>>>> Booleans have a length of 1.
>>>>>
>>>>> Regards,
>>>>>
>>>>> Peter
>>>>>         
>>>> Hi Peter,
>>>>
>>>> You've brought up some interesting points here. Can you actually 
>>>> clarify these changes with regard to parts of the official shapefile 
>>>> spec, or have they been found purely through experimentation?
>>>>
>>>>
>>>> ATB,
>>>>
>>>> Mark.
>>>>
>>>>       
>>> Hi Mark,
>>>
>>> I do still have to figure out, if there are public specs for this. 
>>> All I've found is, that attributes should be in data base files which 
>>> are in dBase format 
>>> (http://www.esri.com/library/whitepapers/pdfs/shapefile.pdf). From 
>>> the analysis of the file (and Wikipedia: 
>>> http://en.wikipedia.org/wiki/Shapefile), it seems that this should be 
>>> dBase III. I do not have any evidence that there is a public spec for 
>>> dBase III.
>>> So all we did is basically gather some experimental data, where 
>>> ArcGIS and gvSIG are our references. The first, because ESRI 
>>> introduced the shape file format and gvSIG, because it is the desktop 
>>> environment that we encourage our customers to use. Anyway, we have 
>>> an old Borland/Inprise tool ("Database Desktop") here for managing 
>>> dBase files and used it for our tests.
>>>
>>> Peter
>>>
>>>     
>>
>>
>>   
> 
> 
> 
> ------------------------------------------------------------------------
> 
> 
> ------------------------------------------------------------------------
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> postgis-users mailing list
> postgis-users at postgis.refractions.net
> http://postgis.refractions.net/mailman/listinfo/postgis-users



More information about the postgis-users mailing list