The ERD File Specification

The ERD file is a standard file format used for storing numbers in tabular form, with labels to support automated plotting and processing.

It was developed by the Engineering Research Division (ERD) of The University of Michigan Transporation Research Institute (UMTRI) and is used in ERD Software.

• Summary
• The Header
• Required Lines
• Optional Lines with Keywords
• The Data Section
• Text Format
• Binary Data

Summary

The ERD file format was developed within the Engineering Research Division (ERD) of the University of Michigan Transportation Research Institute (UMTRI) to facilitate automated plotting of simulation data, experimentally measured data, and data from various analysis programs. A plotter called EP (Engineering Plotter) has been developed for viewing data in ERD files. Versions of EP exist for Mac, Windows, and some UNIX platforms.

An ERD file contains two independent sections, the header and data. The header part contains only text, and the data part contains only numbers. The numbers can be written in either text or binary form. The text form is convenient for viewing and editing data with a word processor, whereas the binary form provides more efficient access for automatic processing. If the data part is in text format, both parts are kept in a single file. However, if the data are in binary format, two files are used. The header part is in an ordinary text file, and the data part is in a file with the name of the text file and the extension .bin. For example, if the header file is named out the name out.bin must be used for the data file. Both files must lie in the same folder.

The Header

The ERD file header consists of a series of conventional text lines that are human readable. These lines contain the information used by post-processing tools to read the numerical data.

Required Lines

As a minimum, the header contains three lines of text. The first line identifies the file as following the ERD format. The second line describes the way that the numerical data are stored in the data section of the file. The third required line is an END statement that indicates the end of the header portion. Any number of optional lines can be included between line #2 and the END line. Table 1 summarizes the lines in an ERD file, and describes the parameters used in line #2 to describe the numerical data.

Listing 1 shows an example header which is fairly brief, consisting of the three required lines and four optional lines. (The optional lines are the ones beginning with the keywords TITLE, SHORTNAM, XLABEL, and XUNITS.)

Listing 1. Short Header for an ERD File with Binary Data.

ERDFILEV2.00
      2,    529,    1,    4232,      1,  1.00000    ,     -1,
TITLE   1993 RPUG Study, Dipstick, Section 1, Measurement 1
SHORTNAMLElev.  RElev.
UNITSNAMft      ft
XLABEL  Distance
XUNITS  ft
END

Looking at the second line of the file shown in Listing 1, we see that the file contains data for 2 channels, with 529 samples/channel, stored as 1 binary record consisting of 4232 bytes, that the data storage format is type 1 (4-byte binary), that the interval between samples is 1.00, and that the status of the auxiliary numbers is -1.

Listing 2 shows a longer header for a file with its data in text form. Note that the data begin immediately after the END line of the header.

Listing 2. Typical Header for an ERD File with Text Data.

ERDFILEV2.00
      2,    529,     -1,     -1,      5,  1.00000    ,     -1,
TITLE   1993 RPUG Study, Dipstick, Section 1, Measurement 1  
SHORTNAMLElev.  RElev.
LONGNAMELeft Elevation                  Right Elevation
UNITSNAMft      ft  
GENNAME Profile Elevation               Profile Elevation
XLABEL  Distance   
XUNITS  ft
FORMAT  (2G14.6) 
PROFINSTDipstick  
HISTORY Converted to ERD format at 23:46, Oct. 23, 1994
END   
  0.000000      0.000000    
  0.416667E-03 -0.141667E-02
  0.416667E-03  0.583333E-03
  0.666667E-03  0.916667E-03
  0.133333E-02  0.133333E-02
  0.750000E-03 -0.166667E-02
 -0.300000E-02 -0.458333E-02
 -0.558333E-02 -0.500000E-02
 -0.625000E-02 -0.658333E-02
 -0.775000E-02 -0.825000E-02

Optional Lines with Keywords

Optional lines in the header begin with an eight-character keyword that defines a particular type of data contained in the remainder of the line. Keywords are associated with one of five general data types: integers, floating point (real) numbers, 8-character names, 32-character names, and 80-character names. The number of data items is either one per file, one per channel, an arbitrary number N, or repeatable.

Table 2 lists common keywords recognized by most post-processing tools. The use of some of these keywords is demonstrated in Listings 1 and 2.

Usually, names associated with a keyword are shorter than the space allowed. When several names are on the same line, the names are padded with blanks as needed so that following names begin at the correct column positions. For example, the header shown in Listing 1 includes names of the units for each channel, as identified with the keyword UNITSNAM. The name of units for the first channel, ft, has only two characters. Thus, it is followed by six spaces so that the name for the second channel, ft, begins in the correct column position.

The Data Section

The data part of the ERD file contains nothing but numbers, organized into columns and rows. The form in which the numbers are stored depends on the value of the KEYNUM parameter from line 2 of the header (see Table 1). The total number of values that will appear in the data section is NCHAN x NSAMP. All of the numbers in the data portion are stored in the same format, and there can be no missing values.

Text Data

The text format is necessary for transporting data in ERD files between different computers, and sometimes even for reading the same file with different programs on the same computer. It is also convenient when numbers are typed in manually, or when numbers are to be edited using a text editor. There are penalties for using text representations of numbers, however. First and foremost, the computer must work hard to translate the text numbers into binary form. It takes about 10 times longer to read a text file than a binary equivalent. A second penalty is that text files take up much more space than binary files.

When data are stored in text form, the numbers are kept in the same file as the header, with the numbers beginning immediately after the header. The ERD file in Listing 2 shows an example of numerical data in text form.

Another option is available when the numbers are always separated by delimiters such as spaces or commas. This occurs when the numbers are obtained by a commercial analysis program or when they are "captured" from another computer. The file of numbers can be made into an ERD file by inserting a 3-line header at the beginning of the file.

If the header of the ERD file does not contain a line with the FORMAT keyword, it is the same as if the FORMAT is a blank. When this occurs for a text file, the file is assumed to contain numbers in free form. The only restriction on free format numbers is that adjacent numbers must be separated. For example, the following line is valid for representing 5 numbers 1.2, 3, 4, -.0201, 14.3:

1.2000 3 4 -2.01E-01 14.3

The following line is not, because the third and fourth numbers touch.

1.2000 3.0000 4.0000-2.01E-01 14.3000

Numbers may be separated by one or more spaces, the tab character, or a comma.

Binary Data

Reading and writing binary data is very efficient, because the computer does not need to perform any conversions or transformations as the data values are moved between the file and the computer memory. When a binary format is used, the data portion of an ERD file is a direct copy of a portion of the computer memory, corresponding to a two-dimensional array having dimensions sized to the number of channels and the number of samples. As indicated in Table 1, two forms of binary data are presently supported: 2-byte integer and 4-byte floating point. 2-byte integer data are typically obtained by data-acquisition systems. Each integer value is a sampled reading obtained from a digitizer during a test. For most engineering applications, data are stored (in the computer memory) in 4-byte floating point format, also known as single-precision floating point. The 4-byte floating point format is commonly used for data generated by computer. The maximum efficiency for data processing is usually obtained when the 4-byte floating point format is used.

The ERDFF is used on a variety of computer systems and for a variety of mass storage media. On some systems, binary data are stored in discrete records. A computer program reading such a file needs to know how many bytes each record contains, and how many records are in the file. Thus, the header contains these two parameters.

Disk files on workstations and desktop computers are not structured: a binary file is simply a continuous stream of bytes that continues to the end of the file. Thus, technically correct parameter values for the header could be one record, containing all of the bytes for the file. Also, there is a certain amount of overhead associated with reading a record. The time needed to read the data for a file is minimized if a single read operation is performed for the entire file.

On the other hand, if the file is large, the memory needed to read the entire file in one chunk may not be available with some programs. A second problem can occur if the true number of bytes in the file is less than the number as inferred by the parameters nbytes and nrecs (i.e., the total size of the file should be nbytes x nrecs). The last "record" is not read, resulting in a loss of data. If the records are large, this loss will could be significant. These problems are reduced if a value of nbytes is specified such that it divides the data into nrecs records of smaller chunks of data.