So, I'm writing a database population script that parses (sic) CSV
files, breaks the input into fields, and then pops it into an RDBMS
(MSSQL, Oracle, Sybase) or Access.
Everything works fine, provided that I my input data is not the exact
length of the destination field while containing non-printable chars.
I have for example, a varchar (255) field, into which I want to put a
string that contains 253 chars (including spaces) and two bell
characters.
The bell characters are interpolated by the apps that use the RDBMSs as
carriage return/new lines. Why not use actual \r\n chars? Uhhmmm..... I
dunno... The apps aren't mine, and they're not going to change just because I
asked them to do so...
When I use any other data loader (BCP, for example) the chars go straight in.
When I use my script, the bell chars get changed, I'm guessing, to Unicode.
Now, I saw that the Text::CSV_XS module has a "binary" switch, and that
DBI::CSV has a csv_csv meta-data parameter. Do I use this to get the "binary"
parameter passed through? Pray tell, how?
Thank you,
amonotod
`\|||/ amonotod@ | subject line:
(@@) charter.net | no perl, no read...
ooO_(_)_Ooo________________________________
_____|_____|_____|_____|_____|_____|_____|_____|
