170121 - AB Export errors

Status: RESOLVED WORKSFORME

(SeaMonkey :: MailNews: Address Book & Contacts, defect)

Description

[Jesse Moorman]

Mozilla 1.0.1 Build ID: 2002082606
W2K
(caveat: I accidentally installed it over 1.0)
(I was aware of the superinstallation issue, but the install process does not
make it perfectly clear whether or not a new subdirectory is being created. As
long as this remains an issue with Mozilla (and some people will not even
realize it is an issue) the version, by default, should be eponymous of the
directory.)

Reproducible every time, but results vary. Some of it may depend on the data.
(Some specific results are unpredictable, some are unpredictable.)
Export Address Book as CSV or Tab delimited
Import/Open in db program.

AB has, according to Mozilla, 538 names. I'm trying to 'import' into Lotus
Approach 9.6 or 'open' the file in Paradox 9 database. The results are a little
different in the two applications. 

It seems that CSV exports are completely fubar: 

If it is a CSV file, in either program, the last time I got 4 columns in
Approach and 2 columns in Paradox, with many items in each columns. Approach
seems to separate each item with tabs, but Paradox reads them as run together.

Using Approach, if I specify that the first line of text file has field names, I
get a table with ~9145 entries and 1 or 2 columns. In one such import, the
column headings were "dn: cn=Jennifer Foster" and "mail=jfoster@c17t" which
(though long) is a truncation of her email address by 10 characters. Many of the
entries in the first column begin with "objectclass: " followed by items like
'top', 'person', 'organizationalPerson', 'inetOrgPerson'.... I got comparable
results in Paradox, but now I cannot remember how I exported that file, Tab of CSV.

Tab Delimited results are less radical.

If I import Tab Delimited into Approach specifing that the first row does not
contain field names, I get ~539 entries and 38 columns, without field names. I
can sort those columns and find a row that is mostly field names, but some
fields do not have names, and when I sort them, I find a few bits of data in
some of those columns, and most of them are entirely empty. In one field (#7), I
get one guy's office fax number, but all other fax numbers are in the field
(#11) that has the fax number label. That guy's info starts out with a perfectly
plausible email address in Field 1 (it has FLast.HisOrganization.org) in the
column where first names go for most people, but this is not an email address
that appears in his Mozilla entry!!! It has his private email (Mozilla's
'Additional Email') in the last name column; his work phone in the email column,
etc...  When I import that (or similar) file into Paradox, that guy's info is in
the correct columns, but the email address has two email address run together,
the second one being the one that is not visible in Mozilla.

Maybe there are relatively few data that are messed up. But how to trust this?

Comment 1

[Jesse Moorman]

I re-installed 1.0.1 clean (2002082606), and get the same kind of results with
CSV export of address book.

Comment 2

[Frank Wein [:mcsmurf]]

Reporter: Is this still a problem with a current Mozilla build?

Comment 3

[Jesse Moorman]

A quick test (Mozilla 1.7b) of two CSV exports, and (after manually changing the
resulting file's extension from .csv to .txt) import into Lotus Approach
(specifying that the first line does NOT containing field names), does not show
any obvious problem. (I'll try other tests later.)

If the bug itself is, indeed, fixed, may I suggest that there is still room for
small enhancements (1) allowing export of field names, (2) in explaining to the
user what the export results will be (e.g. containing field names or not), and
(3) in allowing the '.txt' extension (in lieu of '.csv').

Comment 4

[Frank Wein [:mcsmurf]]

ok resolving then, anyway fill free to file a new bug for this enhancements (but
please only one issue per bug!)