To comment on the following update, log in, then open the issue: http://www.openoffice.org/issues/show_bug.cgi?id=85946
------- Additional comments from [email protected] Thu Mar 25 05:49:12 +0000 2010 ------- I think the first approach should be: 1) For Text Import on Open, use the last used settings for the given filename! I.e. keep a cached mapping with the document history. This may be impractical. 2) Otherwise, detect. The detection of text delimiter should perhaps precede column delimiters, because there is much greater variability in the latter, and it is hard to determine without having first stripped out quoted/escaped portions. Assuming there is no quoting (i.e. it has been stripped, or it does not exist in the first place), one approach to finding column delimiters would be to find all non-alphanumeric characters which have constant frequency >= 1 on all lines of the file. Or if we can't strip quoted text, find all non-alphanumeric, non-quote characters which have a constant minimum value >= 1 on each line. Heuristics or a machine learning approach might then select which result is most appropriate in case of conflict. Alternatively, brute force it: determine which (common?) pairs of delimiters give the text integrity (same number of cols per line), and then use heuristics to decide (e.g. prefer quoted over unquoted; tab over semicolon, comma or space; more columns over fewer?). If OOo could acquire a collection of test text spreadsheets it might be helpful! --------------------------------------------------------------------- Please do not reply to this automatically generated notification from Issue Tracker. Please log onto the website and enter your comments. http://qa.openoffice.org/issue_handling/project_issues.html#notification --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
