Title:
Commercial buildings energy consumption survey. Public use data
diskettes [electronic resource]
Original diskette files: All diskettes were mounted on a 5 1/4" drive
of a Windows 98 PC and the diskette files were copied to the PC hard drive.
To preserve the original file content, all files were transferred to a UNIX
server in binary format and retain their original time and date stamps.
Compressed files:
The 1989 DOS compressed data files were extracted using the file compression
software included on the original diskettes (DOS ARCE.COM) on a Windows 98 PC.
Users should note that Windows OS versions that use the NT (rather than DOS)
kernel will not run the DOS ARCE program. The files were transferred in
binary format with their original time and date stamps.
The 1992 DOS compressed files are Windows self-extracting files. These files
were executed on a Windows PC. All self-extracting decompressed files
were transferred in binary format and retain their original time and date
stamps.
Processed documentation files: The original DOS ASCII documentation files
were examined with Perl "cio" (check-it-out) to identify extraneous control
and high ASCII characters. Form feeds (octal 14) and the substitute control
codes (octal 32) were replaced with a blank using Perl "fix". Octal 376
characters were replaced with a "-" (octal 55) and DOS carriage returns
(octal 15) were deleted with Unix "tr" in the documentation files. The
"fixed" files were saved with .TXT file extensions.
SAS data definitions are part of the original ASCII documentation files.
These data definitions were "cut" from the documentation files and are
available separately as ASCII text files.
Processed data tables and files:
The original 1989 ASCII data files were examined with Perl "cio"
(check-it-out) to identify extraneous control and high ASCII characters.
The substitute control code (octal 32) was replaced with a blank using
Perl "fix". DOS carriage return characters (octal 15) were removed with
UNIX "tr". The "fixed" files were saved with .CSV data file extensions.
The original diskette 1992 ASCII CSV files (BC92*.TXT) are not "fixed" or
translated because these files do not include variable names. The 1992 dBase
data tables, which include variable names, are translated to ASCII CSV format.
Users are advised that the original 1992 ASCII CSV files contain the substitute
control code (octal 32) as the last record or line in the files.
You can view the
1989 and
1992 "cio" output for the original ASCII
diskette files.
The 1992 dBase data tables were translated to ASCII comma separated value
(ASCII CSV) data file format with Windows DBMS/COPY version 7
(www.conceptual.com). Variable names are part of the dBase tables and are
included in the translated files. DBMS/COPY translation logs were checked for data
translation error messages. No comparisons were made on the actual translated
data cell values.
The DBMS data dictionary output statement for file BC92F01D.CSV
follows. Users should note that this statement is specific for this
file only and that each ASCII CSV file has a unique list of variable names
that correspond to each dBase input file.
