Charles Curley wrote: > On Thu, Sep 07, 2006 at 03:43:14PM -0400, R. G. Newbury wrote: >> I have large number of scanned images of documents, which have been >> 'managed' by a windows program which produces an index page with a link >> to each document, and a document page with links to each page in the >> document. There are about 1400 document pages. >> Unfortunately the windows program produces a link in the form: >> <a href='..\index.html'><img src='..\header_index.jpg'> >> That is: containing backslashes. >> >> I have tried to figure our a search/replace line for sed but nothing I >> have tried works...in part I assume because the '\' is itself the >> 'escape' character. Even so, search expressions like '..\\' do not work. >> >> ANyone know how to do this? There are about 20 backslashes in each >> document index page, so I am not even considering a manual method... >> although that works... > > Try this untested line on a throw-away file: > > sed -i s/\\/\//g filename > > I like this form better - you do not have to use / to separate fileds, and I find using @ is more readable in cases like this. I also find that enclosing the regular expression in quotes works better. sed -i -e '[email protected]\\@/@g' filename Mikkel -- Do not meddle in the affairs of dragons, for thou art crunchy and taste good with Ketchup!