Page 235 - Mechanical Engineers' Handbook (Volume 2)

224 Data Acquisition and Display Systems

creating a new variable that reflects some attribute of that relationship. The data collection
system can store the results of that variable, such as the sum of the number of times it
occurred in some larger time interval, allowing it to be low enough in volume to be mass
stored for later analysis.

5.2 File Storage
An easy way to store data is in a file, often a comma-delimited file. This is easy to program
and can easily be imported into analysis tools such as spreadsheets. It is not well suited to
the compression techniques mentioned above because of the complexity of storing and in-
terpreting data. However, for storing records of multiple variables collected at a time interval
this can be a very useful technique. An example is shown in Table 4.
In the above example, the filename was chosen so that it would be unique. The date
and time were concatenated together, eliminating every invalid character with an underscore.
This helps to prevent files from being overwritten accidentally and facilitates store-and-
forward techniques described below. Files can be sorted by date or name. Often, they are
stored in a directory structure so the number of files in any one directory does not get too
great. This speeds up file search in a given directory but can make programs more complex
that search for files across directories. An example directory structure is the following:

C:
DataDirectory
2003
11—files created during November 2003 are stored here
12
2004
01
02
03
Archiving files to backup media is easy in this file organization because one only needs
to reference the particular directory for the time frame desired. Files can fill up mass media
and so either a manual process to check for file limits, back up old files, and delete them
must be instituted or a program to provide the same functions would need to be created.
A major deficiency with a file-based system is that when the time range of a search is
larger than one file, the analysis can become very difficult. One may be searching for events,

Table 4 Example of Storing Multiple Variables in Files
Filename ProcessData20040312142305
Data in file: Timestamp,TemperatureA,TemperatureB,TemperatureC,Rate1,Rate2,Rate3
2004-03-20 12:34,15,14,15,12,13,13
2004-03-20 12:35,15,14,14,12,13,13
2004-03-20 12:36,16,14,13,12,13,13
2004-03-20 12:37,15,14,12,12,13,13
2004-03-20 12:38,15,14,11,12,13,13
2004-03-20 12:39,14,14,11,12,13,13
2004-03-20 12:40,15,14,10,12,13,13
2004-03-20 12:41,15,14,10,12,13,13
2004-03-20 12:42,16,14,10,12,13,13
2004-03-20 12:43,15,14,09,12,13,13