"Kevin Ellis" <email@example.com> wrote in message news:firstname.lastname@example.org... > "james bejon" wrote in message <email@example.com>... >> Thanks very much. I wasn't aware of the 'stable' option. > > Thanks so much James and dpb for all the help. As a sidenote, dpb, you > mentioned that you don't like to use cell arrays and I am curious what the > other options are. Currently, I have a dataset from excel (411,000) rows > of data with 37 columns. Each has either a string or numerical value > depending on the column. I originally had developed a structure array and > imported the data using 'xlsread'. This process took far too long in terms > of runtime so I ended up importing the data using 'xlsread1' and reading > the data directly into a cell array because it was much faster. The > problem is that I work directly with datasets from Excel and have not > found a way I am satisfied with of importing the data to work with into > MatLab. In addition to that, I found structure arrays difficult to work > with on data manipulations such as sorting based on different columns, > removing entries, etc. > So my question is if there is a better way to deal with large datasets > from Excel and import them into a form that can be easily manipulated in > MatLab and output once again to Excel?
If you have the Statistics Toolbox, you can import the Excel data into a dataset. From there it is easy to manipulate the data.