──┬─ FIX ─┬─┬─ = ───────────────┬─────────────────────────────────────
└─ * ───┘ └─<sequence number>─┘
►─┬────────────────────────────────────┬─<edit specs>──────────────────%
└─<start column>─┬───────────────────┤
└─ ─ ──<end column>─┘<edit specs>
──┬──────────┬─ <dlm><target><dlm><new text> ───────────────────┬──────┤ ├─ AFTER ──┤ │ ├─ BEFORE ─┘ │ ├─ INCLUSIVE ─┬─ <dlm><target1><dlm><target2><dlm><new text> ─┘ └─ EXCLUSIVE ─┘
Explanation
The FIX command is used to edit a line of text by inserting new text within the original text, replacing original text with new text, or deleting text.
The basic action of a FIX command is to locate an objective, which can be a character or group of adjacent characters to be deleted or replaced; the objective may also be a point between two characters where an insertion is to be made. When the objective is found, the deletion or insertion is performed. Text to the right of the objective is moved right or left if the insertion is longer or shorter than the deletion; trailing blanks are discarded or inserted as needed. If nonblank characters are pushed past the end of the text field, the FIX operation is performed, but an overflow error is noted.
The asterisk (*) is a synonym for FIX and may be used to edit lines while in single-line sequencing mode. If the asterisk is the first character after the system-supplied sequence number, CANDE scans the line to determine if it is a valid FIX command. If the line is valid, the command is executed; otherwise, the line is entered as a line of text. The asterisk form of the FIX command cannot be used in page-mode sequencing.
<sequence number>
The sequence number of the line to be fixed is specified as an integer or by an equal sign (=); an equal sign refers to the last line entered or fixed and is undefined if no entry or FIX has been entered since the work file was updated.
<start column>
<end column>
The target search may be bounded or supplanted by column specifications; <start column> and <end column> represent integers;<end column> cannot be less than <start column>, and both must lie within the text field for the work file. Column specifications modify the <edit specs> construct as described at the end of the discussion of the <edit specs> construct.
Note that the FIX command differs markedly from the REPLACE command. The FIX command acts on only one line, but the REPLACE command can act on a range of lines. If a line overflow does occur, FIX performs the action but REPLACE does not. (Both situations generate error messages.)
<edit specs>
The <new text> construct can contain any sequence of characters and is null if the end-of-line (EOL) immediately follows the last <dlm>. (In a SCHEDULE session, the end-of-line is assumed to follow the last nonblank character on the line, unless an EOL character has been defined and used.)
When a non-null target is specified, CANDE scans the text field of the specified line in the work file until a sequence of characters is found that exactly matches the target. All characters including blanks are significant. This target is to be replaced by the new text (or deleted if the new text is null) unless BEFORE or AFTER is specified. If BEFORE or AFTER appears, the new text is inserted just preceding or just following the target, respectively.
If the new text is null when BEFORE or AFTER is specified, no change will be made in the line. If the new text is longer than the target text, an overflow of text beyond the right margin may occur, resulting in an error message. If the target is null, the new text is inserted at the beginning of the line of text.
If <target1> and <target2> are specified, and both are non-null, CANDE scans the line of text for a pattern exactly matching <target1>, and then seeks <target2> in the remainder of the text field beyond the<target1>. The new text then replaces the aggregate of the two targets and the text between them (INCLUSIVE) or just the text between them (EXCLUSIVE). If <new text> is null, the line of text does not change. If the new text is shorter than the length of text it is to replace, blanks will replace the remainder of the text bounded by the two targets. If <target1> is null, it is treated by CANDE as the first character in the line of text, and is always included in the text to be replaced, even if EXCLUSIVE is specified. If <target2> is null, it is treated by CANDE as the last character in the line of text, and is always included in the text to be replaced, even if EXCLUSIVE is specified.
If column specifications are included in the FIX command, they modify the <edit specs> construct. If <start column> or <start column>–<end column> is specified, then <target>, <target1> and <target2> are only sought within the inclusive boundaries of <start column> and the end of line, or <start column>–<end column>, respectively. An error message is issued if the specified target or targets are not found within those boundaries. If <target> or <target1> is null, then <start column> takes over their function. If <target2> is null, <end column> functions as <target2>. Column specifications affect determination of the objective but do not limit the text that may be shifted to effect the fix.
If a hexadecimal escape character is set through the CANDE command ESCAPE, then <target>, <target1>, <target2>, and <new text> can contain hexadecimal values delimited by the defined hexadecimal escape character. However, the target delimiter (<dlm>) cannot be the same as the hexadecimal escape character. For more information about the hexadecimal escape character, refer to the ESCAPE command earlier in this section.
The FIX commands are tanked and are executed only when the next UPDATE, LIST, or page-invoking command occurs. Because of tanking, any error messages that result are not displayed until the time of the next UPDATE, LIST, or page-invoking command; therefore, the error messages may not be displayed at the time the command is entered.
The following listing summarizes the many variations of the FIX command. In the following listing, m and n represent integer column numbers; x and y represent text strings. A blank space in the listing denotes an absent <start column>, <end column>, or keyword; a hyphen denotes a null target. Where several variations achieve the same effect, only the simplest variation is shown (degenerate cases like BEFORE or AFTER an empty target are omitted).
The following listing assumes that the new text is not null. If the new text is null, the word replace should be delete, and the word insert should be do nothing.
<start column>
| <end column>
| | <keyword>
| | | <target> or <target1>
| | | | <target2>
| | | | |
| | | | | Summary of Effect
- Insert at beginning.
m - Insert at column <m>.
m n - Replace columns <m> through <n>.
x Replace <target>.
m x Replace <target> found in <m> through end-of-text.
m n x Replace <target> found in <m> through <n>.
B x Insert before <target>.
A x Insert after <target>.
m B x Insert before <target> found in <m>
through end-of-text.
m A x Insert after <target> found in <m>
through end-of-text.
m n B x Insert before <target> found in <m> through <n>.
m n A x Insert after <target> found in <m> through <n>.
I - - Replace entire text field.
m I - - Replace text from column <m> onward.
I - x Replace text through <target2>.
E - x Replace text prior to <target2>.
m I - x Replace from column <m> through <target2>.
m E - x Replace from column <m> until <target2>.
m n I - x Replace from column <m> through bounded <target2>.
m n E - x Replace from column <m> until bounded <target2>.
I x - Replace <target1> and rest of text.
E x - Replace text beyond <target1>.
m I x - Replace bounded <target1> and beyond.
m E x - Replace beyond bounded <target1>.
m n I x - Replace bounded <target1> through column <n>.
m n E x - Replace after bounded <target1> until column <n>.
I x y Replace <target1> through <target2>.
E x y Replace between <targets>.
m I x y Replace as above with
m E x y <targets> bounded in <m> to end of text.
m n I x y Replace as above with
m n E x y <targets> bounded in <m> to <n>.-
The target search in a FIX command is satisfied by the first match encountered. Enough context (or a column restriction) must be provided to unambiguously specify the intended objective.
-
The <dlm> character may not appear in a target, but it may appear in the new text. Inadvertent use of the <dlm> character within the <target> construct is not detected as an error; CANDE processes a shorter <target> and a longer <new text> than is intended.
-
The FIX command makes a distinction between uppercase and lowercase characters. Use the appropriate uppercase and lowercase characters in the <target>, <target1>, <target2>, and <new text> constructs to perform the intended search and insert, replace, or delete operations.
Examples
L 100 ABCDEFGHIJKLMNOPQRSTUVWXYZ 300 IF CARD.PRESENT THEN 400 VIVID HUES OF RED, ORANGE, GREEN AND BLUE 500 X = SQRT (3.1459 * R**4) # FIX 100:FG:HIJ # L = 100 ABCDEHIJHIJKLMNOPQRSTUVWXYZ # FIX = EXCLUSIVE.HIJ.. # L 100 100 ABCDEHIJ #
The following example shows a correct change:
F 300/CARD.PRESENT/BOOLEAN(CARD.AVAILABLE) # L 300 300 IF BOOLEAN(CARD.AVAILABLE) THEN #
The following example shows an incorrect specification for the same change (the period is part of the target text and should not have been used as the delimiter character):
F 300.CARD.PRESENT.BOOLEAN(CARD.AVAILABLE) # L 300 300 IF PRESENT.BOOLEAN(CARD.AVAILABLE).PRESENT THEN # F 400 A.NGE., YELLOW # L 400 400 VIVID HUES OF RED, ORANGE, YELLOW, GREEN AND BLUE # FIX 400 20-50// # L 400 400 VIVID HUES OF RED, # FIX 500 E/(/)/AREA # L 500 500 X = SQRT (AREA) #
The following example shows the use of the asterisk form from within a single-line sequence mode (note that the target text is not uniquely specified, so the first match is used):
M TEST #WORKFILE TEST:SEQ S 100THIS IS A TEST 200* 100/IS/ISNT 300 # L 100 THISNT IS A TEST #
