CHAPTER 6

 

 FORMATTED INPUT/OUTPUT 

 

The emphasis of Chapter 2 was on unformatted input/output. Generally, most programmers use unformatted input and formatted output, with some exceptions. Unformatted input is easier to work with; formatted output has the advantage that it can be made to look exactly as desired. This chapter deals with formatted input/output.


6.1  FORMAT STATEMENT

A FORMAT statement is a nonexecutable statement linked to a READ or WRITE statement, which can be placed anywhere in a program unit after the PROGRAM statement and before the END statement (refer to Fig. 1.1). It consists of a statement label and the keyword FORMAT followed by a format specification enclosed in parentheses.

Both statement label and format specification are required. Within one program unit, two format statements cannot have the same label.

A FORMAT statement has the following form:

label FORMAT(f1, f2, f3, f4, f5, f6, ..., fn)

where label represents the statement label, and f1, f2, ... is the list of field and edit descriptors. These constitute the format specification and must be separated by commas and enclosed in parentheses. For example:

  105 FORMAT(I10,F10.3,E12.5)

A record is a line of input or output, separated from other records by a carriage return. The left parenthesis is the record initiator and the right parenthesis is the record terminator. Both are required. The pair of parentheses causes the transfer of one record according to the format specification enclosed within them.

The slash (/) in a format specification is a record separator, which is both a record terminator and initiator. The presence of a slash causes the current record to terminate and the next record to initiate, thereby increasing the number of records transferred by one.

The record is transferred according to the format specified within the respective initiator-terminator pair. For example,

  105 FORMAT(I10,F10.3/E12.5)

results in the transfer of two records.

  115 FORMAT(/I10,F10.3/E12.5,E13.6/)

results in the transfer of four records, the first and last being empty (null) records. The null record skips one line of input and produces a blank line of output. Hence,

  125 FORMAT()

skips one line, and

  135 FORMAT(///)

skips four lines.

Alternatively, the repeat slash

  135 FORMAT(3(/))

will also skip four lines.

The following are examples of FORMAT statements, described in detail later in this chapter.

  150 FORMAT(A20,I10,F10.3,E12.5/)

  230 FORMAT(//8X,2I5,3F12.5,5(/))

The letters A, I, F, and E are format codes of field descriptors, and X is a format code of an edit descriptor. The comma is required as a field or edit descriptor separator. The comma is also required preceding a repeat slash, as in the last example. The comma is optional when preceding or following one or more slashes, as shown in the following versions of the previous examples:

  150 FORMAT(A20,I10,F10.3,E12.5,/)

  230 FORMAT(//,8X,2I5,3F12.5,5(/))


6.2  FIELD AND EDIT DESCRIPTORS

A field descriptor specifies the size and form of a data item. An edit descriptor affects the location of the data item in the external field. Field and edit descriptors are specified with format codes, which are listed in Table 6.1.

Regarding input, a data item is transferred from an external field (on the screen or in an input file) to an internal value (in the computer's memory). As the programmer, you work with and see only the external field.

Regarding output, a data item is transferred from an internal value to an external field (to the screen or to an output file).


TABLE 6.1  FORMAT CODES

Code Type Function Form

I

F

E

D

L

A

H

X

T

$

Field

Field

Field

Field

Field

Field

Field

Edit

Edit

Edit

Transfers integer values

Transfers real values

Transfers real values with E exponent

Transfers real values with D exponent

Transfers logical values

Transfers character values

Transfers character constants

Specifies character skipping

Specifies positional tabulation

Specifies suppression of carriage return

I w

Fw.d

Ew.d

Dw.d

Lw

Aw

nH...

nX

Tn

$

Integer Field Descriptor I

This field descriptor takes the form

Iw

where I is the format code for integer values and w is the field width, which is the number of spaces allotted. For example:

I5

specifies an integer format occupying up to 5 spaces.

In output, the integer field descriptor has also the form Iw. Optionally, it may take the form

Iw.n

where n is the minimum number of characters that must appear within the field. For example:

I5.3

In this case, assume that 27 is the number to be printed. The output is allowed 5 spaces, and a minimum of three digits must be shown. Since 27 has two digits, an additional leading zero will be printed. Therefore, the output will appear as bb027. (Throughout this chapter, a lower case b represents a blank space).

The value of n should be less than w. If n is greater than w, it will cause an output error. If n is equal to w, it will cause an output error only if the number is negative.

In an input statement, the I field descriptor transfers w characters from an external field into an internal value. The following rules apply (see Integer Input Processing):

  • The external field must contain an integer constant, and it must not have a decimal point or exponential notation.
  • The characters in the external field must be right- justified within the field, i.e., flushed to the right.

  • The internal value is zero if the entire external field is blank.
  • Leading zeros or leading blank spaces in the external field are ignored. Conversely, trailing zeros or trailing blank spaces in the external field are taken as significant zeros and are made part of the internal value.

Integer Input Processing

Field Descriptor

External Field

Internal Value
I3

I5

I4

I6

I10

bb

12345

b-35

bb-20b

bbbb357246

0

12345

-35

-200

357246

Formatting of output is much more common than formating of input. Upon output, the I field descriptor transfers an internal value to an external field of w characters, right-justified. The following rules apply (see Integer Output Processing):

  • If the internal value does not fill completely the external field, leading blank spaces are inserted.

  • If the internal value does not fit within the external field because it is too large, the external field overflows (a write overflow) and is entirely filled with asterisks (*).

  • If the internal value is negative, the external field must be large enough to accomodate the negative sign.

  • If n is present, the external field consists of n digits (not including the sign) and is filled with leading zeros if necessary.

The following examples show integer output processing, from internal value to external field.


Integer Output Processing

Field Descriptor Internal Value

External Field

I3

I4

I5

I5

I6

I8

I5.1

I5.3

I5.5

I5.5

-123

-35

12345

135246

-200

-357246

0

32

4365

-789

***

b-35

12345

*****

bb-200

b-357246

bbbb0

bb032

04365

*****


Example: Integer Field Descriptor

C234567890


      READ(5,110) I,J,K


  110 FORMAT(I3,I5,I8)


      WRITE(6,120) I,J,K


  120 FORMAT(I5,I8,I10)

This example reads integers I, J, and K from a file connected to unit 5, according to format 110. Assuming that the values are:

I= 5, J= 23, K= 367

The input (external field) should look like this:

bb5bbb23bbbbb367

The output (external field) is then written according to format 120, and would look like this:

bbb5bbbbbb23bbbbbbb367

Note that the first character of output was not printed. This is because the first column of output is never printed, and used instead for the control of carriage return. Specific rules regarding carriage control will be introduced later in this chapter.

Things to watch for in formatted input/output of integer data:

  • With input, remember to right-justify the integer values in the external field. Gross errors may occur if you do not follow this rule.

  • With output, make sure to specify an external field large enough to accomodate the largest number. Otherwise, an overflow will occur and the output will instead contain a set of asterisks.


Real Field Descriptor F

This field descriptor takes the form

Fw.d

where F is the format code for real values, w is the field width, and d is the number of characters to the right of the decimal point. For example:

F7.2

specifies a real format occupying up to 7 spaces, of which 2 are to the right of the decimal point.

The (optional) sign and decimal point in a real number are counted as one space each. For example, -135.79 would occupy 7 spaces.

In an input statement, the F field descriptor transfers w characters from an external field into an internal value. The following rules apply (see Real Input Processing on the next page):

  • The internal value is positive if the first nonblank character of the external field is a plus sign, or if no sign appears. The internal value is negative if the first non blank character of the external field is a minus sign.
  • The internal value is zero if the external field contains only blank spaces or a decimal point (see Section A below).
  • If the external field includes a decimal point: (a) the location of that decimal point overrides the location speci fied by the d; and (b) leading and/or trailing blanks are filled with zeros and are not significant (see Section B below).
  • If the external field includes neither a decimal point nor an exponent: (a) leading and/or trailing blanks are filled with zeros; (b) the field is treated as a real value of w characters; and (c) d is meaningful, with the rightmost d digits taken as the fractional part of the real number (see Section C below).
  • If the external field includes an exponent, i..e, the letter E followed by an optional exponent sign and number, that exponent is used to establish the magnitude of the internal value (see Section D below). In this case, the characters in the external field should be right-justified . The following examples illustrate real input processing (transfer from external field to internal value).


Real Input Processing


Section Field Descriptor External Field Internal Value
A F5.1

F5.1

bbbbb

bb.bb

0

0

B F6.2

F8.3

F10.3

1.2345

12345.67

b54.3bbbbb

1.2345

12345.67

54.3

C F6.2

F6.2

F10.2

F10.2

123456

b1234b

bbb-12345.

-12345bbbb

1234.56

123.40

-12345.

-1234500.00

D F10.3

F13.6

F13.6

F13.6

bb1275E-01

bb123.456E-10

bbb1234567E-7

bb1234567E-07

1.275E-1

0.123456E-07

0.1234567E-06

0.1234567E-06


More commonly used is real output formatting, where the F field descriptor transfers w characters from an internal value to an external field of w characters, rounded to d decimal positions and right-justified. The following rules apply:

  • If the internal value does not fill the entire external field, leading blanks are inserted in the external field.

  • If the internal value does not fit within the external field, the field overflows and is entirely filled with asterisks (*).

Note that the character field width w must exceed d by at least 3 in order to fit in the (optional) sign, the decimal point, and at least one digit to the left of the decimal point.


Real Output Processing


Field Descriptor Internal Value External Field

F4.1

F4.1

F5.1

F5.2

F8.3

F8.3

F8.5

F8.5

F10.4

F10.6

F10.8

15.632

1.63

-32.54

-32.54

1234.56

1234.567

12.34567

12345.67

-12.34567

-12.34567

-12.34567

15.6

b1.6

-32.5

*****

1234.560

1234.567

12.34567

********

bb-12.3457

-12.345670

**********


Example: Real Field Descriptor

C234567890
      READ(5,210) A,B


  210 FORMAT(F10.3 F12,5)


      WRITE(6,210) A,B


  220 FORMAT(F12.3, F15.5)

This example reads real variables A and B from a file connected to unit 5, according to format 210. Assuming that the values are:

A= 123.654, B= 45.89

The input (external field) should look like this:

123.654bbb45.89bbbbbbb

The output (external field) is then written according to format 220, and would look like this:

bbbb123.654bbbbbbb45.89000

Again, notice that the first character of output was not printed (control of carriage return).

Things to keep in mind in formatted input/ output of real data using the F field descriptor:

  • With input, always use a decimal point to ex plicitly state its location and avoid having to count columns when entering data.
  • With output, make the field width w large enough to avoid a write overflow. If unsure of how large, use the E field descriptor instead.


Real Field Descriptor E

This field descriptor takes the form

Ew.d

where E is the format code for exponential notation, w is the field width, and d is the number of characters to the right of the decimal point. For example:

E13.6

reserves 13 spaces, 6 of which are to the right of the decimal point. The signs, the letter E, and the decimal point are counted as one character each (for example, the number -1.23E-02 takes up 9 spaces).

Input and output operations for the E field descriptor work in the same way as the F field descriptor (Section A). However, w must exceed d by at least 5 (but better yet 7, see Section B) in order to fit, from left to right: the decimal point, the letter E, an exponent sign, a 2-digit exponent, an optional sign, and a possible zero to the left of the decimal point (the latter may be omitted by the processor if necessary to avoid a write overflow, see Section C).


Real Output Processing with E Field Descriptor

Field Descriptor Internal Value >>>>> External Field
A E10.2

E10.3

E12.7

0.246357

1234567.

-1234.567

bb0.25E+00

b0.123E+07

************

B E12.5 -0.0001234 -0.12340E-03
C E12.6

E 12.7

-1234.567

1234.567

-.123457E+04

.1234567E+04


Example: Real Field Descriptor E

C234567890

READ(5,310) C,D

310 FORMAT(E10.3,E15.5)

WRITE(6,320) C,D

320 FORMAT(E12.3,E15.5)

This example reads real variables C and D from a file connected to unit 5, according to format 310. Assuming that the values are:

C= -12.3, D= -45.89

The input (external field) should look like this:

-0.123E+02bbb-0.45890E+02

The output (external field) is then written according to for mat 320, and would look like this:

b-0.123E+02bbb-0.45890E+02

Again, notice that the first character of output was not printed (control of carriage return).


Real Field Descriptor D

The D field descriptor is used for double-precision vari ables. This field descriptor takes the form

Dw.d

where D is the format code, w is the field width, and d is the number of characters to the right of the decimal point. For example:

D15.8

reserves 15 spaces, 8 of which are to the right of the deci mal point. Input and output operations for the D field de scriptor work in the same way as the E field descriptor, ex cept that the letter D appears instead.


Logical Field Descriptor L

This field descriptor takes the form

Lw

where L is the format code for logical values and w is the field width. For example:

L7

reserves 7 spaces.

In an input statement, the L field descriptor transfers w characters from an external field into an internal value. The stored value is either .TRUE. or .FALSE.

The following rules apply:

  • The value is .TRUE. if the external field starts with a T, or with a T preceded by a period (.T).
  • The value is .FALSE. if the external field starts with an F, with an F preceded by a period (.F), or is blank.
  • If none of the above, an input error occurs during execu tion.
  • If w= 1, the external values must be either T, F, or blank.
  • If w= 2, the external values must be either .T, .F, start with a T or F, or be blank.
  • If w is omitted, a value w=2 is assumed.

In an output statement, the L field descriptor transfers the letter T if the internal value is .TRUE. or the letter F if the internal value is .FALSE. The value is right- justified in the external field. If w is omitted, a value is assumed (usually w= 2).

The following examples illustrate logical input/output processing.


Logical Input Processing

Field Descriptor External Value >>>>> Internal Field

L5

L7

L9

L

L

L2

L2

.TRUE.

FRANKbb

bbbbbbbbb

T

b

TR

FL

TRUE

FALSE

FALSE

TRUE

FALSE

TRUE

FALSE



Logical Output Processing

Field Descriptor Internal Value >>>>> External Field
L5

L7

L

TRUE

FALSE

TRUE

bbbbT

bbbbbbF

bT


Example: Logical Field Descriptor L

C234567890

LOGICAL P,Q

READ(5,410) P,Q

410 FORMAT(L5,L5)

WRITE(6,420) P,Q

420 FORMAT(L3,L3)

Assuming that P is .TRUE. and Q is .FALSE, the input should look like this:

TRUEbFALSE The corresponding output would be:

bTbbFb


Character Field Descriptor A

This field descriptor takes the form

Aw

where A is the character format code and w is the field width.

For example:

A10

will allot 10 spaces for a character variable.

In an input statement, the A field descriptor transfers w characters from an external field into an internal character value of length v.

The character length v is specified by a CHARACTER declaration statement. For example:

CHARACTER*12

where v= 12.

The following rules apply:

  • If w is equal to v, w = v characters are transferred.
  • If w is less than v, w characters are transferred, left- justified, and as many trailing blanks are added as nec essary to fill the length v.
  • If w is greater than v, the rightmost v characters are transferred and the leftmost excess characters are lost.
  • If w is omitted, v characters are transferred. If the exter nal value has z characters, where z < v, the internal field is filled with z characters and v - z trailing blanks. If z > w, the internal field is filled with the leftmost v charac ters.


Character Input Processing
Field Descriptor Character Length External Field >> Internal Value

A6

A6

A8

A

A

6

8

6

8

6

ABCDEF

ABCDEFGH

ABCDEFGH

ABCDEF

ABCDEFGH

ABCDEF

ABCDEFbb

CDEFGH

ABCDEFbb

ABCDEF

In an output statement, the A field descriptor transfers w characters from an internal value of length v to an exter nal field of length w. The following rules apply:

  • If w is equal to v, w = v characters are transferred.
  • If w is less than v, only the leftmost w characters are transferred.
  • If w is greater than v, v characters are transferred, right- justified, with leading blank spaces.
  • If w is omitted, v characters are transferred. If the inter nal value has z characters, where z < v, the external field is filled with z characters and v - z trailing blanks. If z > v, the external field is filled with the leftmost v characters.

Character Output Processing

Field Descriptor Character Length External Field >> Internal Value

A6

A6

A8

A

A

6

8

6

8

6

ABCDEF

ABCDEFGH

ABCDEF

ABCDEF

ABCDEFGH

ABCDEF

ABCDEF

bbABCDEF

ABCDEFbb

ABCDEF


Example: Character Field Descriptor A

C234567890

CHARACTER WORD*10,LABEL*12,NAME*15

READ(5,510) WORD,LABEL,NAME

510 FORMAT(A10,A12,A)

WRITE(6,520) WORD,LABEL,NAME

520 FORMAT(1X,A10,A12,A)

This example reads real variables WORD, LABEL, and NAME from a file connected to unit 5, according to format 510. Assuming that the values are:

WORD= 'YELLOW', LABEL= 'TURQUOISE',

NAME= 'KIRKPATRICK'

The input (external field) should look like this:

YELLOWbbbbTURQUOISEbbbKIRKPATRICKbbbb

The output (external field) is then written according to for mat 520, and would look like this:

YELLOWbbbbTURQUOISEbbbKIRKPATRICKbbbb

In this example, a 1X edit descriptor (See Edit Descrip tor X and Carriage Control later in this section) is in cluded as the first edit descriptor listed in the output FOR MAT statement (with label 520). If the 1X is omitted, the first character of variable WORD (in this case, Y) is used for carriage control and is not printed.

Things to keep in mind in formatted input/output of character data:

  • Try to match the field width w with the charac ter length v to avoid loss of data or added blank spaces.

  • Consider omitting the field width entirely. In this case, the specified character length will be used as field width.


Character Field Descriptor H

This field descriptor takes the form

nHabc & def

where H is the format code and n specifies the exact num ber of characters to be transferred, to follow immediately after the H code (in this case, the characters abc & def). Leading, embedded, and trailing blank spaces are included in the n count.

For example:

17H VELOCITY PROFILE

will transfer 17 characters, including blanks.

A limitation of the H field descriptor is that the number of characters following H has to equal n exactly. Begin ning with Fortran 77, there are other more convenient ways of writing character constants, which avoid the counting of characters. Therefore, the H field descriptor has been categorized as obsolescent by the Fortran 90 standard.

As an alternative to the A and H field descriptors, char acter constants may be specified directly; either leading, trailing, or embedded within (1) variable lists in WRITE statements, or (2) field and edit descriptor lists in FOR MAT statements (See Example on next page).

In most of these examples, a 1X edit descriptor is used to account for carriage control (See Edit Descriptor X and Carriage Control later in this section). If 1X is omitted (as in format 620), a blank space leading the character con stant can be used for carriage control (' THE VALUE OF B IS = ').


Example: Use of Character Constants

C234567890

WRITE(6,610) A

610 FORMAT(1X,20HTHE VALUE OF A IS = ,F10.3)

WRITE(6,620) ' THE VALUE OF B IS = ',B

620 FORMAT(F10.3)

WRITE(6,630)'THE VALUE OF C IS = ',C,' UNITS'

630 FORMAT(1X,A,F10.3,A)

WRITE(6,640) D

640 FORMAT(1X,'THE VALUE OF D IS = ',F10.5, 1' UNITS')

WRITE(6,650) Y,Z

650 FORMAT(1X,'Y= ',F10.3/1X,'Z= ',F10.5)

Assuming A = -23.4562, B = 123456.789, C= 75.9, D= -123.45678, Y= -12345.032, and Z = 9876.5432; the output would look like this:

THE VALUE OF A IS = -23.456

THE VALUE OF B IS = 123456.789

THE VALUE OF C IS = 75.900 UNITS

THE VALUE OF D IS = -123.45678 UNITS

Y= -12345.032

Z= 9876.54321


Edit Descriptor X

This edit descriptor takes the form

nX

where X is the format code and n specifies the number of characters that are passed over, or jumped. The value of n must be greater than or equal to 1. If n= 1, it may be op tional in some processors. However, to ensure compatibil ity among processors, it is good practice to always precede the X code with a number, even if it happens to be 1.

Several X edit descriptors within one format specifica tion result in the skipping of a combined number of spaces, as in the following example:

20 FORMAT(1X,'LENGTH',5X,'WIDTH',5X,'HEIGHT')

which results in

LENGTHbbbbbWIDTHbbbbbHEIGHT


Edit Descriptor T

This edit descriptor takes the form

Tn

where T is the format code and n specifies the number of characters that are tabulated; n must be greater than or equal to 1. In an input statement, the first n characters of the record are skipped. In an output statement, Tn tabs to column n of the record (column n-1 of the printed output, since the first column of the record is never printed).

Several T edit descriptors within one format specifica tion result in positional tabulation, each time from the beginning of the record, as in the following example:

20 FORMAT(T21,'COLUMN 2',T11,'COLUMN 1')

which results in

C23456789012345678901234567890

COLUMN 1 COLUMN 2

In this example, the character constant COLUMN 1 is printed starting in column 11 of the record (column 10 of the printed output, since the first column of the record is never printed). The character constant COLUMN 2 is printed starting in column 21 of the record (column 20 of the printed output).

Edit Descriptor $

The edit descriptor $ (optional with some processors) modifies the carriage control specified by the first charac ter of the record. This descriptor can be placed anywhere in the list of descriptors. If more than one $ sign appears, the first one is used and the others are ignored.

In an input format, the edit descriptor $ is ignored. In an output format, the $ descriptor suppresses the carriage return if the first character of the record is a blank space. This feature is convenient when using direct screen input (interactive input), because the input data record can be entered on the same line as the input prompt. This im proves the readability of the input data being entered to the screen.


Example: Edit Descriptor $

C234567890

WRITE(6,700)

700 FORMAT(1X,'ENTER VALUE OF A: ',$)

READ(5,*) A

This example allows the value of A to be typed on the same line as the input prompt. As the input prompt is

ENTER VALUE OF A:

the value of A, say 12.3, is to be entered to the screen im mediately following the input prompt. The prompt and in put line on the screen would look like this:

ENTER VALUE OF A: 12.3

Carriage Control

With output, the first character of a record is reserved to specify the number of carriage returns (carriage control). Table 6.2 lists valid carriage control characters; b stands for a blank space or for any other character not listed in this table. If a field descriptor for carriage control is omit ted, either by design or involuntarily, the first character of the record is used for carriage control. If this happens, that character is taken over by the processor and not printed.

TABLE 6.2

CARRIAGE CONTROL CHARACTERS

Character No. of Returns Meaning
b

0

1

+

$

1

2

several

0

0

Go to next line (single spacing)

Skip one line (double spacing)

Skip to top of next page (paging)

Go to the beginning of the same

line (overwrite previous record)

Suppress single-spacing carriage

return (remain at the end of the line).

It is good programming practice to explicitly account for the intended number of carriage returns at the begin ning of each output format. Otherwise, there is the possi bility of unpredictable output. For instance, if the first character of every line happens to be a 1, one page will be used per line of output.


Example: Carriage Control

C234567890

WRITE(6,810) A

WRITE(6,820) B

WRITE(6,830)

WRITE(6,840)

WRITE(6,850)

WRITE(6,860)

WRITE(6,870)

810 FORMAT(1X,'THE VALUE OF A IS = ',F10.3)

820 FORMAT(' THE VALUE OF B IS = ',F10.3)

830 FORMAT(' ','GO TO NEXT LINE, SINGLE SPACING')

840 FORMAT('-','GO TO NEXT LINE, SINGLE SPACING')

850 FORMAT('0','SKIP ONE LINE, DOUBLE SPACING')

860 FORMAT('1','SKIP TO TOP OF NEXT PAGE')

870 FORMAT('+','GO TO BEGINNING OF THE SAME LINE.',

1'WRITE OVER PREVIOUS RECORD')


Repeat Count of Field Descriptors

The field descriptors I, F, E, D, L, and A can be repeated within one record by preceding any of them by an integer, intended as a repeat count.

For example, the following formats are equivalent:

10 FORMAT(I5,I5,F10.4,F10.4,F10.4)

20 FORMAT(2I5,3F10.4)

Several field descriptors can be repeated as a group, by enclosing the list in parentheses and preceding it by an integer, intended as a group repeat count. If a group re peat count is omitted, a value of 1 is assumed.

These formats are equivalent:

30 FORMAT(1X,I3,E12.5,I3,E12.5/)

40 FORMAT(1X,2(I3,E12.5)/)


Output Statement Record Overflow

With output, up to 133 columns can be specified per record, including the first column, which is reserved for carriage control. For example, these two statements are equivalent ways to print a series of 132 dashes:

WRITE(6,100)

100 FORMAT(133(H-))

WRITE(6,200)

200 FORMAT(1X,132(H-))

On the other hand, the following format statement will cause a record overflow (an execution error):

100 FORMAT(1X,133(H-))

This format specifies 134 columns, one more than the maximum of 133 allowed.


6.3 OTHER FORMAT TOPICS

Interaction of Input/Output Statements with Format Statements

In Fortran, the input/output (READ and WRITE) and format statements interact in a specific way to allow the orderly input/output of a list of variables. This interaction is described in this section.

There is a close relationship between input/output variable lists and their respective field descriptor lists specified in the format. The basic concept to remember is that control remains with the READ or WRITE statement. A READ or WRITE statement executes the transfer of a number of data items equal to the number of variables listed in the statement.

  • If the field descriptor list is longer than the variable list, the latter takes precedence. From left to right, the processor uses only as many field descriptors as needed to accomplish the transfer of all items listed on the vari able list. The remaining field descriptors are ignored.

  • If the field descriptor list is shorter than the variable list, the processor treats the format specification as typi cal, and repeats it as many times as needed to accom plish the transfer of all items on the variable list. During the repeated execution of the field descriptor list, the processor ignores any remaining field descriptors.

The following examples illustrate the interaction of in put/output and format statements.

Example 1

READ(5,100) A,B,I,J

100 FORMAT(F10.3,F10.4,I5,I10)

The variable list in the READ statement matches in number and type the field descriptor list in the FORMAT statement. Two real and two integer values will be read.

Example 2

READ(5,200) A,B,C,D

200 FORMAT(F10.3,F10.4,F10.5,F10.6)

The variable list matches in number and type the corre sponding field descriptor list. Four real values (A, B, C, and D) will be read in sequence.

Example 3

READ(5,300) A,B

300 FORMAT(F10.3,F10.4,F10.5,F10.6)

The variable list is shorter than the corresponding field descriptor list. Only the first two field descriptors will be used, and two real values (A and B) will be read.

Example 4

WRITE(6,400) A,B,C,D

400 FORMAT(1X,F10.3)

The variable list is longer than the corresponding field descriptor list. The format is interpreted as typical. All four variables will be written, one per line, following the field descriptor specified in the format. The output may look like this:

12345.678

-98.765

-0.345

-98765.432

Example 5

WRITE(6,500) A,B,C,D

500 FORMAT(1X,F10.3,F10.4/1X,F10.5,F10.6)

The variable list matches the corresponding field descriptor list in number and type. The slash in the format indicates that two records (lines) will be written. In total, four real values will be written (A, B, C, and D), two per line. The output may look like this:

12345.678 -98.7654

-0.34543 -8.112345

Example 6

WRITE(6,600) A,B,C

600 FORMAT(1X,F10.3/1X,F10.4/1X,F10.5,F10.6)

 The variable list is shorter than the corresponding field descriptor list. Only the first three field descriptors will be executed. Their types match those of the variable list. The presence of two slashes in the format indicates that three records (lines) will be written. In total, three real val ues will be written, one per line. The output may look like this:

12345.678

-98.7654

-0.34543

Example 7

WRITE(6,700) A,B,C,D,E,F,G

700 FORMAT(F10.3/F10.4/F10.5,F10.6)

The variable list is longer than the corresponding field descriptor list. The format is interpreted as typical. All seven variables will be written following the typical for mat. The second time, only the first three field descriptors will be executed. Their types match those of the variable list. The presence of two slashes in the format indicates that three records will be written in the first pass, and three on the second. In total, seven real values (A, B, C, D, E, F, and G) will be written from six records (A in the first re cord, B in the second, C and D in the third, E in the fourth, F in the fifth, and G in the sixth).

The output may look like this:

12345.789

9876.5432

-106.34521 -2.987654

-14567.897

-2230.6784

9876.43245

Things to keep in mind when matching variable lists and field descriptor lists:

  • Avoid using mixed-type variable and field descriptor lists, particularly with long integer and real lists.

  • If you use mixed-type lists, make sure that the listed items agree in number and have match ing data types. Any mismatch will cause an execution error. This is a very common source of error.

  • Single-type variable and field descriptor lists may disagree in number of elements. How ever, this will either disregard a part of the format, or will transfer more records than the number specified in the format.

Runtime Formats

Some processors allow the READ or WRITE statement to vary its format during program execution. This is referred to as a "runtime format." For this purpose, a character con stant or character variable is used in a READ or WRITE statement instead of the fixed format statement label. When using a character variable in a runtime format, make sure that it is defined in length and value prior to the in voking READ or WRITE statement.

A runtime format frees the input/output statement from having to refer to a format label. In certain cases, the for mat statement may be somewhere else within the program unit, far from the invoking READ or WRITE statement.

The following examples illustrate runtime formats. For comparison, Example 1 shows the conventional paired WRITE and FORMAT statements, where the format is fixed. Example 2 uses a character constant containing the appropriate format specification. Example 3 defines a character variable FMT and uses it in the FORMAT state ment. All three examples are equivalent, and cause the same variables to be output in exactly the same way.

Example 1: Conventional Fixed Format

WRITE(6,100) A,B,C

100 FORMAT(1X,3F10.3)

Example 2: Runtime Format (Character Constant)

WRITE(6,'(1X,3F10.3)') A,B,C

Example 3: Runtime Format (Character Variable)

CHARACTER*11 FMT

DATA FMT /'(1X,3F10.3)'/

WRITE(6,FMT) A,B,C

Variable Formats

Some processors allow a WRITE statement to vary its format specification during program execution. This is re ferred to as a "variable format." For this purpose, a valid arithmetic integer expression, enclosed in angle brackets <>, is used to replace an integer constant wherever it is used as part of a format specification.

Example Program: Variable Formatting

C234567890

PROGRAM VAR_FOR

DATA J,K,L,M,N /3,5,7,9,0/

10 N= N+1

IF(N.LE.8) THEN

WRITE(6,100) J,K,L,M

100 FORMAT(1X,4I)

GO TO 10

ENDIF

END

In this program:

  • The digits 3, 5, 7, 9, and 0 have been initialized in a DATA statement.

  • A logical IF/GO TO combination results in the execu tion of a WRITE statement 8 times.

  • The character field width of format 100 is incremented by one (1) every time the WRITE statement is executed. The output would look like this:

3 5 7 9

3 5 7 9

3 5 7 9

3 5 7 9

3 5 7 9

3 5 7 9

3 5 7 9

3 5 7 9


6.4 EXAMPLE PROGRAMS

Example Program 1: Table of Trigonometric Functions

C234567890

PROGRAM TRIG_FUNCTIONS
CHARACTER*11
1ANGLE*7,SINE,COSINE,TANGENT,COTANGENT,SECANT,COSECANT,INFINITY
CHARACTER*19 LINE
DATA PI,IANGLE /3.1415926,-5/
DATA
1ANGLE,SINE,COSINE,TANGENT,COTANGENT,SECANT,COSECANT,INFINITY
2 /' ANGLE', ' SINE',' COSINE',' TANGENT',
3 ' COTANGENT',' SECANT',' COSECANT',' INFINITY'/
DATA LINE /'-------------------'/
WRITE(6,500) LINE,LINE,LINE,LINE
WRITE(6,100) ANGLE,SINE,COSINE,TANGENT,COTANGENT,SECANT,COSECANT
WRITE(6,500) LINE,LINE,LINE,LINE
20CONTINUE
IANGLE= IANGLE + 5
X= FLOAT(IANGLE)*PI/180.
SINX= SIN(X)
COSX= COS(X) IF(IANGLE.EQ.0) THEN
TANX= TAN(X)
SECX= 1/COSX
WRITE(6,200) IANGLE,SINX,COSX,TANX,INFINITY,SECX,INFINITY
ELSEIF(IANGLE.LT.90) THEN
TANX= TAN(X)
COTX= 1/TANX
SECX= 1/COSX
CSCX= 1/SINX
WRITE(6,300) IANGLE,SINX,COSX,TANX,COTX,SECX,CSCX
ELSEIF(IANGLE.EQ.90) THEN
COTX= 1/TANX
CSCX= 1/SINX
WRITE(6,400) IANGLE,SINX,COSX,INFINITY,COTX,INFINITY,CSCX
ELSE
WRITE(6,500) LINE,LINE,LINE,LINE
STOP
ENDIF
GO TO 20
100FORMAT(1X,A,1X,6A)
200FORMAT(1X,I5,3X,3F11.6,A,F11.6,A)
300FORMAT(1X,I5,3X,6F11.6)
400FORMAT(1X,I5,3X,2F11.6,A,F11.6,A,F11.6)
500 FORMAT(1X,4A)
END

This program shows how to calculate and print a table of values. The program calculates and prints all six trigo nometric functions for angles varying between 0o and 90o, at increments of 5o.

Note the following features of this program:

  • The integer IANGLE has been initialized at -5 to enable the calculation of the angle to start at 0o.

  • The block IF construct allows the computation and printing of the special cases (0o and 90o).

  • The character constant INFINITY is printed where the value of the function is ¥.

  • The character constant LINE has been used to draw separating lines. The output of program TRIG_FUNCTIONS is shown below.

-----------------------------------------------------------------------------------------------------

ANGLE SINE COSINE TANGENT COTANGENT SECANT COSECANT

-----------------------------------------------------------------------------------------------------

0 0.000000 1.000000 0.000000 INFINITY 1.000000 INFINITY

5 0.087156 0.996195 0.087489 11.430053 1.003820 11.473714

10 0.173648 0.984808 0.176327 5.671281 1.015427 5.758770

15 0.258819 0.965926 0.267949 3.732051 1.035276 3.863704

20 0.342020 0.939693 0.363970 2.747478 1.064178 2.923805

25 0.422618 0.906308 0.466308 2.144507 1.103378 2.366202

30 0.500000 0.866025 0.577350 1.732051 1.154701 2.000000

35 0.573576 0.819152 0.700207 1.428148 1.220775 1.743447

40 0.642788 0.766044 0.839100 1.191754 1.305407 1.555724

45 0.707107 0.707107 1.000000 1.000000 1.414213 1.414214

50 0.766044 0.642788 1.191753 0.839100 1.555724 1.305407

55 0.819152 0.573577 1.428148 0.700208 1.743447 1.220775

60 0.866025 0.500000 1.732051 0.577350 2.000000 1.154701

65 0.906308 0.422618 2.144506 0.466308 2.366201 1.103378

70 0.939693 0.342020 2.747476 0.363970 2.923803 1.064178

75 0.965926 0.258819 3.732049 0.267949 3.863701 1.035276

80 0.984808 0.173648 5.671280 0.176327 5.758769 1.015427

85 0.996195 0.087156 11.430044 0.087489 11.473705 1.003820

90 1.000000 0.000000 INFINITY 0.087489 INFINITY 1.000000

---------------------------------------------------------------------------------------------------

Example Program 2: Variable Formats

C234567890

CHARACTER X2*2

DATA L,X2 /0,'XX'/

10 L= L+1

IF(L.LE.11) THEN

WRITE(6,100) X2,X2

GO TO 10

ENDIF

100 FORMAT

*(<1+ABS(5-ABS(6-L))>X,A,X,A)

END

This program uses a variable format to print the char acter pattern shown below. (Note that your com piler/processor may not have this feature).

XX XX

XX XX

XX XX

XX XX

XX XX

XXXX

XX XX

XX XX

XX XX

XX XX

XX XX

The features of this program are:

  • L varies from L = 1 to L = 11. One line is printed per value of L. L = 1 for line 1; L = 2 for line 2; and so on.

  • The intrinsic function ABS (see Section 5.4) is used to calculate the variable number of blank spaces to be in serted between the XX's.

  • The carriage control character (1X) has been incorpo rated into the first field descriptor listed in Format 100.


6.5 SUMMARY

This chapter describes formatted input/output by linking a FORMAT statement to a READ or WRITE statement. The FORMAT statement contains field and edit descrip tors. The field descriptors I, F, E, D, L, A, and H specify the size and form of the data item. The edit descriptors X, T, and $ affect the position or location of the data item. Format codes are used to specify field and edit descriptors.

READ or WRITE statements interact with FORMAT statements in a specific way to allow the orderly in put/output of a long list of variables.


CHAPTER 6--PROBLEMS

  1. Write a program to calculate and print a table of hyperbolic trigonomet ric functions, for angles between 0 and 6 radians, in increments of 0.1 radians. Use six decimal figures.

  2. Write a program to calculate and print a table of decimal logarithms, for values ranging from 1.00 to 9.99, in increments of 0.01. Arrange the ta ble in double-entry form, as follows:

    x 0 1 2 3 4 5 6 7 8 9

    __________________________________________________________

    1.0

    2.0

    3.0

    etc.

    Use four decimal figures.

  3. Write a program to calculate and print a table of natural logarithms, for values ranging from 1.00 to 9.99, in increments of 0.01. Arrange the ta ble in double-entry form (see Problem 2). Use four decimal figures.

  4. Do Problem 4.4 (baggage handling) using formatted input/output.

  5. Do Problem 4.5 (water utility billing) using formatted input/output. Set output in tabular form, in four columns labeled NAME, ADDRESS, CONSUMPTION, and BILLING.

  6. Do Problem 4.10 (electric utility billing) using formatted input/output. Set output in tabular form, in four columns labeled NAME, ADDRESS, CONSUMPTION, and BILLING.

  7. Write a program to calculate and print a table of the hyperbolic sine, for values ranging from 0.00 to 5.99, in increments of 0.01. Arrange the ta ble in double-entry form (see Problem 2). Use two decimal figures.

  8. Write a program to calculate and print a table of the hyperbolic cosine, for values ranging from 0.00 to 5.99, in increments of 0.01. Arrange the table in double-entry form (see Problem 2). Use two decimal figures.

  9. Write a program to calculate and print a two-column table. The left col umn should have integer N, varying from 0 to 30, and the right column its factorial. Use the E field descriptor to print the factorial.

  10. Write a program to calculate and print a two-column table. The left col umn should have integer N, varying from 1 to 100, and the right column its cubic root, with 6 decimal figures.

  11. Write a program to calculate and print a two-column table. The left col umn should show ANGLE (RADIANS), varying from 0 to 10, in inter vals of 0.1 radians. The right column should show ANGLE (DEGREES, MINUTES, SECONDS), the latter with 1 decimal figure, as follows: 6.7 radians (corresponding to) 383o 52' 54.2" Use double-precision variables.

  12. Write a program to print the following design using a variable format:

    X X X X X

    X X X X X X X X

    X X X X X X X X

    X X X X X X X X

    X X X X X X X X

    X X X X

  13. Write a program to print the following diamond design using a variable format:

    DD DD DD DD

    DD DD DD DD

    DD DD DD DD

    DD DD DD DD

    DD DD DD DD

    DD DD DD DD

    DD DD DD DD

    DD DD DD DD

    DD DD DD DD

    DD DD DD DD

    DD DD DD DD

  14. Write a program to print the following 3 letters (XYZ) using a variable format (as much as possible) and 5 blank spaces between letters.

    X X Y Y ZZZZZZZZZZZ

    X X Y Y Z

    X X Y Y Z

    X X Y Y Z

    X X Y Y Z

    X Y Z

    X X Y Z

    X X Y Z

    X X Y Z

    X X Y Z

    X X Y ZZZZZZZZZZZ