Class DataStoring currently stores data in memory. Furthermore it provides a method to store data "manually" to a file.
In extension to this an optional automatic data export to the file system is proposed. This is sensefull when
a) The amount of data exceed the memory capacities
b) A loss of data in case of a process breakdown is not tolerable
The following buffer types are proposed:
Everything
All incoming data rows are buffered. No automated file export -> current state of implementation
Nothing
A new data row is appended to the associated file immediately.
Current frame
Only the current frame is buffered. At the beginning of a new frame or if the data storing object is destroyed, the last frame is appended to the associated file.
Number of rows
A defined number of rows is buffered.
New parameters for the constructor could be:
Buffer type (default: Everything)
Optional destination path (mandatory for all buffer types except 'Everything')
Optional file name stub (w/o suffix; mandatory for all buffer types except 'Everything')
The final file name is generated at actual storage time like this: yyyy-mm-dd hh:mm:ss.msec [file name stub].csv
Optional maximum number of files (default = 0, means: no file limit)
If >0 the oldest files are deleted automatically
With this extension we can realize something like a "flight recorder".
Class DataStoring currently stores data in memory. Furthermore it provides a method to store data "manually" to a file.
In extension to this an optional automatic data export to the file system is proposed. This is sensefull when a) The amount of data exceed the memory capacities b) A loss of data in case of a process breakdown is not tolerable
The following buffer types are proposed:
New parameters for the constructor could be:
With this extension we can realize something like a "flight recorder".
To be discussed...