by In an earlier version of Ingres (was it some version of 5 or 6??) We
copied data in from an ASCII file using the following code:
COPY TABLE uo_sample(log_no='c0|',
lab_loc_id='c0|',
requester='c0|',
depth='c0|'
,sampled='c0|',
note='c0|',
lab='c0|',
doc_cntrl_no='c0|',
matrix='c0')
FROM 'sample.xfer';
The table is defined as:
Key
Column Name Type Length Nulls Defaults Seq
log_no c 12 no yes
loc_id c 12 no yes
sampled date yes null
norm_year integer 2 no yes
norm_qtr integer 1 no yes
norm_month integer 1 no yes
type c 4 no yes
depth float 4 no yes
project c 4 no yes
footnote c 5 no yes
note c 150 no yes
lab c 10 no yes
requester c 20 no yes
lab_loc_id c 30 no yes
dup_spike c 10 no yes
entered date yes null
updater c 3 no yes
doc_cntrl_no c 6 no yes
matrix c 2 no yes
time_zone c 3 no yes
validation c 1 no yes
labcode c 2 no yes
last_updated date no yes
regulator c 1 no yes
wgmgok c 1 no yes
If one of the columns in one of the records of the ASCII file was too
big we would receive the following message:
COPY: Warning: 1 character domains were truncated when copied from the
file to character type columns in the table. (E_CO0015)
Since it was a warning, the copy would complete with one field in one
column getting truncated data.
Now, in OpenIngres1.1, we recieve this message, and the copy FAILS. We
want the behavior of the first case where we get a warning and the copy
completes with a truncation.
Does anyone know a way around this?
Francesca Colombini
Lawrence Livermore National Laboratory