csvgraph 4v0

Csvgraph now works with Unicode text (uft-8) so any characters can be used. That means that a csv file that is uft-8 encoded will be correctly read, and its BOM (Byte Order Mark) if present correctly processed. Column headings, the main titles and axis titles can all include utf8 characters. Utf-8 is backwards compatible with 7-bit ASCII so “conventional” csv files should be processed exactly as before. When saving a csv file, you have the option to specify that the file is created with a BOM. Filenames and the paths to files may also contain Unicode characters (utf-8). Note that saving a CSV file with recent versions of Excel creates a utf-8 format csv file with a BOM, and a BOM is required for Excel to recognise the csv file if you try and open it.

csvgraph 3v10

Added Smoothing spline filter

CSVgraph 3v9

Changes with 3v9:

Can now open a file that excel already has open (and error messages are better on failing to open files).

Better trapping of user pressing a "command" button while a previous command is still running.

Derivative now uses 17-point Savitzky Golay algorithm with user specified order (1->10 is actually used, can be set 1->infinity by user).

Savitzky Golay smoothing added as a filtering option (25 points, with user specified order, 1->10 is actually used, can be set 1->infinity by user).

Added 2nd derivative (d2y/d2x) to list of filters which uses a 25-point Savitzky Golay algorithm with user specified order (1->10 is actually used, can be set 1->infinity by user).

If a number is missing in a column referred to in an expression this will be set to nan.

Added constant "nan" for expressions.

If an expression evaluates to nan the line is skipped so this can be a powerful way to select points for csvgraph to display.

Added "variables" x and line to expressions. x is current x value, and line is current line number.

Updated expression handler so nan==nan and nan!=nan work as expected in expressions.

csvgraph 3v7

3v7 10/6/2023 –
“Show legends” tick-box added.
Changed to using Builder C++ 11.3 compiler.
Title is now centred above the graph.
Minor changes to sizes, fonts, etc in csvgraph.

csvgraph 3v6

3v6 3/6/2023
Changes from 3v5:

• Long column headers now cause a scroll bar to be automatically added to the X & Y listboxes so they can be fully seen.
• Save x range on screen as CSV added to File menu.
• Option (tickbox) added to add basename of filename to legends of traces on the graph, which is useful if the same column is read from multiple files.
• Y axis title automatically added unless user specifies one (based on column header of 1st trace added).
• Added option to load X as Value/60 (sec->min), Val/3600 (sec->hrs), val/86400 (sec->days).
• Error handling for X values in a user defined date/time format improved, and trailing whitespace now allowed.

csvgraph 3v5

On a right mouse click when a line is selected then the slope of this line (dy/dx) is given as well as the coordinates of both ends of the line.

A filename can be given on the command line (on earlier versions this did not work for the 64-bit version).

An expression containing a function is now allowed for the y axis (e.g. max($2,0) ).

csvgraph 3v4

This fixes a dpi handling issue with multiple monitors on the 64 bit version.
No changes to functionality.

csvgraph 3v3

The 64 bit version of this release has been optimised for loading very large csv files. With 16GB of RAM a 100GB file can be loaded and viewed (this uses almost all Windows default allocation of virtual memory). Note that using virtual memory slows the program down, but this is mainly when loading the file, zooming is still very fast.

cscgraph 3v2

1st 64-bit release.

64-bit version will use all available memory, while 32-bit version is limited to 4GB of ram, apart from that functionality is identical. The 64-bit version can read files with more than 2^32 lines, the 32-bit version will run out of RAM before that many lines can be read.

Minor changes:

• When the Scales menu was invoked multiple times previously the scaling would change very slightly – fixed.
• Allowed range of font sizes for main title and X/Y axis titles expanded.
• FFT now makes use of multiple processor cores if they are present to deliver faster results.

csvgraph 2v9

2v9 – 7/6/2022 – bug fix – if x-offset is not equal to 0 and multiple traces added incorrect x-offset
was applied to the 2nd trace added onwards.
Higher resolution internally on reading dates & times, so if “start time from zero” is
ticked the results may be slightly more accurate.