Running ODA Software on Windows: Assigning Missing Values using ...

Optimal Data Analysis Vol. 6 (January 1, 2017), 3-4

Copyright 2017 by Optimal Data Analysis, LLC 2155-0182/10/$3.00

Running ODA Software on Windows: Assigning Missing Values using Excel and using Longer File Names Timothy A. Ebert, Ph.D., and Paul R. Yarnold, Ph.D. University of Florida, Department of Entomology and Nematology, Citrus Research and Education Center, 700 Experiment Station Rd., Lake Alfred

Optimal Data Analysis, LLC

For users running ODA software using Windows 7-10 operating systems we describe how to assign missing values via Excel, and we note that file names may consist of more than eight characters.

satisfac.dat;” (the ODA command in prior Windows systems) may instead now be “open satisfaction.dat;”. For the input data file, the file name is limited to 81 characters. This must include the entire path and the extension. There must be an extension but it can be any 1, 2, or 3 alphanumeric characters. We also tested the “$” character as the extension “.x$p” and that worked, however we did not test other special characters. The program name can be over 108 characters. At that point the file name exceeded the width of our window and we risked making errors by going longer. The output file can be 125 characters long. This must include the path name (if needed), and the extension (if any). File names longer than this do not generate any errors. However, MegaODA only uses the first 125 characters and truncates the rest. The special characters ` ~ ! @ # $ % ^ & ( ) _ - + ‘ : ? < > , and = are all valid, but using * | \ “ or / results in an error. Using a semi-colon (;) in the output file name does not work, but does not result in an error message. Using a period works

Preparing a data set for analysis by ODA (i.e., UniODA, MegaODA, CTA) software is an exacting process—data must conform exactly to structure defined by the input command.1-4 In this note we present tips useful in constructing and in naming data sets in the Windows 7-10 operation systems. Some popular software systems used for data manipulation represent a missing value on a variable for an observation as a single period (“.”). Periods and other missing value indicators can easily be changed to any desired value if the data exist within, or can be easily imported into Excel. For example, to change missing values indicated as “.” to “-9”: (a) select all cells and hit ctrl-h (find and replace); (b) enter “.” into find; (c) enter “-9” into replace; (d) click on the options button and select “entire cell contents”; (e) click “replace all”; and (f) resave. For Windows 7-10 operating systems, the file names in ODA programs are no longer limited to eight or fewer characters. Thus, “open 3

Optimal Data Analysis Vol. 6 (January 1, 2017), 3-4

Copyright 2017 by Optimal Data Analysis, LLC 2155-0182/10/$3.00

2

but this is the delimiter for the file extension. We suspect that there are a few other special characters that will not work, but most will not cause a problem. Users should be aware that using the full range of possibilities available in modern Windows™ operating systems, as allowed by ODA™ software, may cause problems elsewhere. For example, the cloud storage used by the University of Florida will not sync the file names that contain # or % characters. So, not all programs will be as flexible in terms of characters allowed in file names.

Yarnold PR, Soltysik RC (2016). Maximizing predictive accuracy. Chicago, IL: ODA Books. DOI: 10.13140/RG.2.1.1368.3286 3

Bryant FB, Harrison PR (2013). How to create an ASCII input data file for UniODA and CTA software (Invited). Optimal Data Analysis, 2, 2-6. 4

Ebert TA (2016). Getting started with ODA software: A short tutorial (Invited). Optimal Data Analysis, 6, 235-240. Author’s Notes

References

IRB review was not required for this programming note, and no conflict of interest was reported.

1

Yarnold PR, Soltysik RC (2005). Optimal data analysis: A guidebook with software for Windows. Washington, DC, APA Books.

4