File Formats



File Formats

INPUT FILES:

1. COLS.TXT

COLS.TXT is an ASCII text file containing the column numbers of the raw USPS cost matrix file which are needed in the PRC roll forward program.

Each record consists of one field that represents a column number. This number is in columns 1-3.

The list of column numbers is developed from Docket No. R97-1, USPS LR-H-4, "Base Year Roll Forward, Input Data Files" at 45-46. That document defines a set of 3-digit numbers (between 101 and 200, inclusive) that correspond to mail categories, special services, totals, and subtotals. Most column numbers corresponding to totals or to subtotals were removed from this list to produce the set of column numbers contained in COLS.TXT.

2. COMP.TXT

COMP.TXT is an ASCII text file that contains the list of PRC and USPS component numbers required as input to the PRC roll forward program. The source of these component numbers is file PRCCOMP.XLS of Docket No. MC96-3, PRC-LR-5.

This file also contains two fields used to back out the manual input file from the input base year cost file. One field indicates whether a PRC component number should have attributable costs set to zero (MINPUT) and the other (NOFIXED) indicates whether the component should be 100% attributable (no fixed cost).

|Columns |Name |Description |

|1-4 |PRCN |PRC component number |

|6-10 |ROWN |USPS component number |

|11 |MINPUT |=1, if attrib costs to be zeroed |

| | |=0, otherwise |

|13 |NOFIXED |=1, if fixed costs should be set to zero |

| | |=0, otherwise |

PRCN is the PRC component number and ROWN is the row number of the input USPS cost matrix corresponding to that value of PRCN.

The variable ROWN may be missing. This means that an input record for this value of PRCN is required by the roll forward program, but the cost data for this value of PRCN does not come directly from any particular row of the USPS cost matrix.

Only PRCN values < 2191 are required in this file. There are some larger values on the PRCCOMP.XLS file (Docket No. MC96-3, PRC-LR-5), but they are not needed as input to the PRC cost model.

It is possible for one USPS component number to be associated with two PRC component numbers. In other words, there may be two records having the same values for ROWN but different values for PRCN.

3. B.DAT

This is the USPS cost matrix file in Docket No. R97-1. For Docket No. R97-1, B.DAT was included in USPS library reference H-6. B.DAT is an ASCII text file containing a record for each USPS cost component and field for each category of mail.

Number of records: 1600

Variables per record: 200

Variable type: character

Characters per record: 2205

The record format for this file is as follows:

Columns Variable Length Description

2 VARO 14 USPS component number

17 VAR1 10 Cost data for category 1

28 VAR2 10 Cost data for category 2

…

6+11k VARk 10 Cost data for category k

…

6+11*199 VAR199 10 Cost data for category 199

The values in VAR0,…, VAR199 are zero-filled from the left and include a character appended to the right. The appended character indicates the unit's place and the sign of the data value contained in the VARk field.

The character variables can be converted to integers as follows:

1. Remove leading zeros.

2. If the final character is "A",…,"I" or "{" then the value is positive. If the final character is "A",…,"I" then the alpha character should be changed to "1",…,"9", respectively. If the final character is "{", then it should be changed to "0" .

3. If the final character is "J",…,"R" or "}", then the value is negative and a negative sign should be inserted as the leftmost character. If the final character is "J",…,"R", then it should be changed to "1",…, "9", respectively. If the final character is "}", it should be changed to "0".

4. The resulting character string should consist of the characters "0", …, "9", and possibly "-" as the leftmost character. This character string should be converted to an integer.

Example: If VAR101="000011386D", it should be decoded as 113864 because D is the fourth letter in the alphabet.

If VAR101="000011386R", then it decodes as -113869.

If VAR101="000011386}", then it decodes as -113860.

If VAR101="000011386{", then it decodes as 113860.

4. PREPROCk.SAS

This is the SAS program that creates the manual input cost matrix from the cost matrix provided in the Postal Service's filing. The program requires a cost matrix and the COLS.TXT and COMP.TXT files as input. The logic of the program is documented in the comments embedded in the SAS program.

The program's input statement for "DATA A" should only require modification if the Postal Service's cost matrix is provided in a different format, or if the number of columns in the matrix is changed from 200.

Changes in the number of components or mail categories should be reflected in the COMP.TXT and COLS.TXT input files described above. The program should not require further modifications to reflect changes to input files.

Exception: If the number of mail categories drops so that the value of “&ncol” is less than 33, then the data step that outputs the final data file should be modified. For example, if &ncol=28, then "DATA F" should be modified to read:

data f; set e;

array x{&ncol} x1-x&ncol;

file out;

do k=1 to 16;

put x{k} @;

end;

put;

do k=17 to &ncol;

put x{k} @;

end;

put prcn;

(DATA G should be modified similarly.)

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download