benito
12th February 2003, 22:53
This text appear on every sequential dump of table that I create,
2003-02-11 17:59:55.562
#$!pre-dump!$#00017cv
I observed this after I upgraded to porting set 6.1c.06.04.
Anybody has seen this one before? How do you turn this off?
Thanks a lot in advance.
victor_cleto
13th February 2003, 20:58
You can't turn it off, check the portingset guide or changes, this was implemented to avoid problems where the field splitter appears also in a field, causing the data to corrupt while importing.
From that portingset on, the header includes not only the export date but also information about what is the field splitter. The only downturn is that brokes compatability with previous portingset versions.
nick_rogers
13th February 2003, 22:19
yes I ran into this new feature when exporting data to another Baan installation that has an older porting set. After a few hours of pulling my hair out - the only way to make the older porting set version of bdbpost work is to remove the fist 2 lines of the dump file (created by the newer port set).
NPRao
13th February 2003, 23:19
fyi... if you remove the 1st 2 lines, this wouldnt work in the newer version of the porting sets.
>bdbpost6.2 -c100 -t"|" -A -D .
Table M Idx RowCount Messages
----------- - --- -------- ----------------------
Timestamp incorrect or failing for './tzmmig103100.S' (dump ignored)
tzmmig103100.S open error 0
benito
17th February 2003, 22:34
Below is baan's reply. By changing the bdbpre version in my bin directory i can still go back to the old functionality (no header texts). currently i have 2 versions in this folder, i turn one on depending on the purpose of my dump file.
These extra lines are there since bdbpre has the new format.
This is since the following portingsets:
BaanIV: 6.1c.06.01
BaanERP 5.0c: 7.1b.01
BaanERP 5.0b: 7.1c.02 (the first 7.1c porting set for 5.0b)
Here is more explanation about this new format:
The first line is a timestamp and should comply to the format
"0000-00-00 00:00:00.000" (the character '0' stands for a digit).
The second line is a "magic" ("#$!pre-dump!$#") followed by the version
number of the dump followed by the options that describe the dump.
The version number is encoded as a 4 digit decimal number ("0001").
The separator is encoded as a 2 digit hexadecimal number ("00").
The last character 'v' stands for 'variable format', as opposed to 'f'
which stands for 'fixed format' (-x option).
Further the precision in which floats are dumped is also increased. Doubles and floats are now dumped with maximum precision
This modification is made since the following portingsets:
BaanIV: 6.1c.06.01
BaanERP 5.0c: 7.1b.02
BaanERP 5.0b: 7.1c.02 (the first 7.1c porting set for 5.0b)
Doubles and floats are now dumped with maximum precision. The format of
a float or double in a pre-dump is:
<sign><mantissa>'e'<sign><exponent>
If the format of the dump is fixed (-x option specified), then the <mantissa>
of doubles contains 17 digits and the <exponent> contains 3 digits. Ex.
-12345678901234567e+123. For floats the <mantissa> contains 9 digits and the
<exponent> contains 2 digits. Ex. -123456789e+12.
If the format of the dump is not fixed (-x option not specified), then the
<sign>'s are omitted if they are '+'. If the <exponent> consists of '0's
only, it is omitted, including the leading 'e' and <sign>. Furthermore,
trailing '0's are removed from the <mantissa>. If the 'plain' notation
(without an <exponent>, but including a decimal point) happens to be shorter,
the 'plain' notation is used.
Examples of doubles in non-fixed format are:
-12345678901234567e123 (<sign> before <exponent> is omitted)
123 (<sign> and <exponent> are omitted)
-1.23 ('plain' notation is shorter than -123e-2)
123e-8 (<sign> is omitted)
.0123 ('plain' notation is shorter than 123e-4)
Further bdbpost is made compatible with the old format (so without the 2 lines at the beginning).
This modification is there since the following portingsets:
BaanIV: 6.1c.06.01
BaanERP 5.0c: 7.1b.02
BaanERP 5.0b: 7.1c.02 (the first 7.1c porting set for 5.0b)