reading text file a matrix of string

Hi, i have a large data file in which each row looks like this:

1/1/2012 00:05:00 0.45

all rows have the same format. i want to read it into matlab by using the command "tdfread". it works, but the first 2 columns comes out wrong, the third one is ok.

is there any way to make the command read the first two columns as a string/char and not as a scalar?

using "fopen" takes too long, thats why im trying to use this command.

thank you for your help.

[dpb]

So fast but wrong would be better than right but slow???

What is the delimiter and does the file have a header row that is allowable variable names?

Looks like if those conditions were followed tdfread should do the job. I doubt it'll be any faster than would be textscan or some of the alternatives, however...

[Dany]

i added a header row: date time height

and the delimiter is ' ' (space).

still doesnt work.

i have a text file with at least 105000 rows of data. what do you think would be the faster way to read it and then place the readed data in a matrix.

when i used "fopen" and "fgetl" it took several minutes, its preatty long dont you think?

[dpb]

doc tdfread

"... tdfread(FILENAME,DELIMITER) uses the specified delimiter in place of tabs.

doc textscan  % alternatively.

Using fgetl is one record at a time; about as slow a technique as you could choose.

[Dany]

ok thank you. i'll try the "textscan" command, maybe it'll work batter then the "tdfread"

[dpb]

Quite likely-- tdfread is an m-file that sucks in the whole file as stream and then tries to figure out what it is that it has read.

textscan otoh is a builtin function that uses a specific format string and the i/o libraries...

[Dany]

it works perfect. thank you