We are now gathering CSV-files from our customers, what normally are exports from other systems, to test your driver.
We found a few interesting numbers, which in fact aren't numbers, but strings: telephone numbers, ip-adresses and personnel numbers.
id;telephone1;telephone2;telephone3;telephone4;telephone5;telephone6;personnelnumber;ipaddress
1;0152700900;(015)2700900;015-2700900;015 2 700 900;+31(15)2700900;+31152700900;000345;192.168.1.13
Your driver type-guesses the columns telephone1, telephone6, personnelnumber and ipaddress as a INTEGER, BIGINT, INTEGER and a DOUBLE with this url:
jdbc:text:/.?delayedClose=0&_CSV_Separator=\\x3b&csvfileExtension=csv&_CSV_Quoter=&maxScanRows=0&_CSV_Header=true&charSet=ISO-8859-1&decimalSeparator=,&groupingSeparator=.
Is there a way of detecting these values as a string? For instance: a number usually doesn't start with a "0" or a "+".
|
>Your driver type-guesses the columns telephone1, telephone6, personnelnumber
>and ipaddress as a INTEGER, BIGINT, INTEGER and a DOUBLE with this url:
>Is there a way of detecting these values as a string? For instance: a number
>usually doesn't start with a "0" or a "+".
The latest package support most of them. But +31152700900 will be thought as a bigint type. It will be availabe in 2 hours.
|
Cool, this works nicely and will make it easy to use. This was one of the things our consultants always have to change, when using the Microsoft ODBC Driver.
That the detection of +31152700900 fails isn't that bad, because you won't loose too much information, usually only the +.
|
Main
webmaster@hxtt.com