Ah yes, I see that I have to move the point into the table cell. I was trying with the table header. Slightly odd that. Means that it only works on tables that aggregate clock times across multiple files, where the times are put in the same cell. Can you replicate? If you do a clocktable with the scope set to that file, then there's no way to order the cells. How hard would it be to modify org-dblock-write do you think? In hours work for someone familiar with elisp, but not the org codebase. On 31 March 2014 03:06, Nick Dokos wrote: > Noah Slater writes: > > > Yeah, tried that. Doesn't work! :( > > > > AFAICT, it works fine on your first stackoverflow example. > > There is probably no hope of getting this method to work the way you > want on your second example though: org-sort does not know anything > about the substructure of the table. The only way I can think of is to > make the dynblock function that produces the table > (org-dblock-write:clocktable) do the sorting. > > > > On 30 March 2014 23:24, Nick Dokos wrote: > > > > Noah Slater writes: > > > > > I posted a question on StackOverflow: > > > > > > > http://stackoverflow.com/questions/22749704/how-can-you-sort-an-org-clock-table > > > > > > Summary is: how do I sort an clock table by the % column? > > > > > > Is there anything "out there" I can use to get this working? If > not, > > > how complex a job would it be to write something that did this? > > > > > > If you point me in the right direction, I'll see what I can come > up with. > > > > > > > Never tried on a clock table, but the following works on a generic > > table, so I assume that it will work on a clock table too: put point > > in the column by which you want to sort the table (in the body of the > > table, not in the header) and say M-x org-sort RET n (I assume you > > want numeric sorting, but org-sort provides several kinds). org-sort > > is normally bound to C-c ^ too, so > > > > C-c ^ n > > > > should be all that's needed. > > -- > > Nick > > > > -- > Nick > > >