• choose Save as ... from the file menu.
• Set the Save as Type to be CSV (Comma delimited).
• Save the file.

From within Gsharp:

• Click on the open icon
• Set the Files of type field to be CSV (*.csv)
• Select the file.
• If you wish to specify the number of header lines, whether the file is rows, columns or a grid and whether it should be read as floats or strings then click on the Options button.
• Click on OK.

The report reader will examine the lines before the first data line to find strings that it can use as names for the datasets it will create. If it cannot find an appropriate line then it will use T1, T2, T3, ... for any float columns and A1, A2, ... for any string columns. Strings must be delimited by two spaces. Floats can be separated by a single space.

It is also possible to replace specified strings in the file before it is read using a conversion filter. The filter has the format "/find/replace/". You can include multiple filters e.g. "/find/replace/ /find/replace/". Examples of conversion filters include:

• To convert (34) into -34 use  "/(/-/ /)//"
• To convert "1,2,3,4" into "1 2 3 4" use "/,/ /"

The conversion filter and the name of the titles dataset can be set by clicking on the Options button.

The report reader is useful in that it can handle different numbers of columns but if you know the exact number of columns then the ASCII reader is quicker and uses less memory.

• You must specify either the number of datasets or the names of the datasets, e.g. "6" or "x1 x2 x3 y1 y2 y3"
• You must specify whether your data is floats or strings.
• If you are reading a single dataset then you can use the rows, columns and planes to inform Gsharp of the dimension of the data. You could also use the GSL command reshape once the data has been read. N.B. The layout resource has no affect on how single datasets are read.
• If you are reading multiple datasets then you must inform Gsharp whether the datasets are arranged in columns (vertical) or in rows (horizontal).
• You can also tell the reader only to look for data within a certain window in the file using the options start row, end row, start column and end column. N.B. In this context a column is a character column not a column of data.
• You can use a conversion string as described in the report reader section

N.B. The ASCII reader completely ignores line feeds. It works through the file looking for numbers wherever they are and then stores the values it finds into datasets as instructed by you.

It is also possible to read the whole file as an array of strings and then use GSL commands such as mask, strvalue and slicex to create your datasets.

There is no reason why you can't making multiple calls to import_ascii to read data from the same file.

Some examples:

All text is ignored
1  2
3
4

Data for London
1    2    8    
Data for Birmingham
4   3   2  

Data for London
1    2    8
Data for York
4   3   2

1  2
3  4
5  6

grid = transpose(grid);

• The format of each binary number: float, byte, int or short.
• The name of the dataset to create.
• The dimension of the dataset to create (rows, columns and planes).

If you have problems reading your binary file:

• Experiment with the different number formats until you are getting meaningful numbers.
• Read the whole file without specifying any dimension.
• Use slicex to remove any unwanted numbers.
• If necessary, use reshape to convert your numbers into a grid or block.

Folders are most useful for archiving your data, so that you can come back and use it again in a later session.