使用 COBOL，如果指定名称的文件不存在，如何创建行顺序 .dat 文件？
使用其中之一
OPEN OUTPUT fd
每次都会创建一个新文件擦除任何现有数据，或者将OPTIONAL关键字添加到FILE-CONTROL SELECT短语中。
ENVIRONMENT DIVISION.
INPUT-OUTPUT SECTION.
FILE-CONTROL.
select OPTIONAL data-file
assign to file-name
organization is line sequential
file status is data-file-status.
并使用OPEN I-O数据文件。 如果需要，COBOL 将为第一次写入准备一个空文件。
请参阅 http://opencobol.add1tocobol.com/#does-opencobol-support-isam 获取任意示例（不是按行顺序，但应该明白要点）。
出于完整性考虑，此处包含列表（和输出）：
OCOBOL >>SOURCE FORMAT IS FIXED
*> ***************************************************************
*><* ================
*><* indexing example
*><* ================
*><* :Author: Brian Tiffin
*><* :Date: 17-Feb-2009
*><* :Purpose: Fun with Indexed IO routines
*><* :Tectonics: cobc -x indexing.cob
*> ***************************************************************
identification division.
program-id. indexing.
environment division.
configuration section.
input-output section.
file-control.
select optional indexing
assign to "indexing.dat"
organization is indexed
access mode is dynamic
record key is keyfield of indexing-record
alternate record key is splitkey of indexing-record
with duplicates
.
*> ** OpenCOBOL does not yet support split keys **
*> alternate record key is newkey
*> source is first-part of indexing-record
*> last-part of indexing-record
*> with duplicates
data division.
file section.
fd indexing.
01 indexing-record.
03 keyfield pic x(8).
03 splitkey.
05 first-part pic 99.
05 middle-part pic x.
05 last-part pic 99.
03 data-part pic x(54).
working-storage section.
01 display-record.
03 filler pic x(4) value spaces.
03 keyfield pic x(8).
03 filler pic xx value spaces.
03 splitkey.
05 first-part pic z9.
05 filler pic x value space.
05 middle-part pic x.
05 filler pic xx value all "+".
05 last-part pic z9.
03 filler pic x(4) value all "-".
03 data-part pic x(54).
*> control break
01 oldkey pic 99x99.
*> In a real app this should well be two separate flags
01 control-flag pic x.
88 no-more-duplicates value high-value
when set to false is low-value.
88 no-more-records value high-value
when set to false is low-value.
*> ***************************************************************
procedure division.
*> Open optional index file for read write
open i-o indexing
*> populate a sample database
move "1234567800a01some 12345678 data here" to indexing-record
perform write-indexing-record
move "8765432100a01some 87654321 data here" to indexing-record
perform write-indexing-record
move "1234876500a01some 12348765 data here" to indexing-record
perform write-indexing-record
move "8765123400a01some 87651234 data here" to indexing-record
perform write-indexing-record
move "1234567900b02some 12345679 data here" to indexing-record
perform write-indexing-record
move "9765432100b02some 97654321 data here" to indexing-record
perform write-indexing-record
move "1234976500b02some 12349765 data here" to indexing-record
perform write-indexing-record
move "9765123400b02some 97651234 data here" to indexing-record
perform write-indexing-record
move "1234568900c13some 12345689 data here" to indexing-record
perform write-indexing-record
move "9865432100c13some 98654321 data here" to indexing-record
perform write-indexing-record
move "1234986500c13some 12349865 data here" to indexing-record
perform write-indexing-record
move "9865123400c13some 98651234 data here" to indexing-record
perform write-indexing-record
*> close it ... not necessary, but for the example
close indexing
*> clear the record space for this example
move spaces to indexing-record
*> open the data file again
open i-o indexing
*> read all the duplicate 00b02 keys
move 00 to first-part of indexing-record
move "b" to middle-part of indexing-record
move 02 to last-part of indexing-record
*> using read key and then next key / last key compare
set no-more-duplicates to false
perform read-indexing-record
perform read-next-record
until no-more-duplicates
*> read by key of reference ... the cool stuff
move 00 to first-part of indexing-record
move "a" to middle-part of indexing-record
move 02 to last-part of indexing-record
*> using start and read next
set no-more-records to false
perform start-at-key
perform read-next-by-key
until no-more-records
*> read by primary key of reference
move "87654321" to keyfield of indexing-record
*>
set no-more-records to false
perform start-prime-key
perform read-previous-by-key
until no-more-records
*> and with that we are done with indexing sample
close indexing
goback.
*> ***************************************************************
*><* Write paragraph
write-indexing-record.
write indexing-record
invalid key
display
"rewrite key: " keyfield of indexing-record
end-display
rewrite indexing-record
invalid key
display
"really bad key: "
keyfield of indexing-record
end-display
end-rewrite
end-write
.
*><* read by alternate key paragraph
read-indexing-record.
display "Reading: " splitkey of indexing-record end-display
read indexing key is splitkey of indexing-record
invalid key
display
"bad read key: " splitkey of indexing-record
end-display
set no-more-duplicates to true
end-read
.
*><* read next sequential paragraph
read-next-record.
move corresponding indexing-record to display-record
display display-record end-display
move splitkey of indexing-record to oldkey
read indexing next record
at end set no-more-duplicates to true
not at end
if oldkey not equal splitkey of indexing-record
set no-more-duplicates to true
end-if
end-read
.
*><* start primary key of reference paragraph
start-prime-key.
display "Prime < " keyfield of indexing-record end-display
start indexing
key is less than
keyfield of indexing-record
invalid key
display
"bad start: " keyfield of indexing-record
end-display
set no-more-records to true
not invalid key
read indexing previous record
at end set no-more-records to true
end-read
end-start
.
*><* read previous by key of reference paragraph
read-previous-by-key.
move corresponding indexing-record to display-record
display display-record end-display
read indexing previous record
at end set no-more-records to true
end-read
.
*><* start alternate key of reference paragraph
start-at-key.
display "Seeking >= " splitkey of indexing-record end-display
start indexing
key is greater than or equal to
splitkey of indexing-record
invalid key
display
"bad start: " splitkey of indexing-record
end-display
set no-more-records to true
not invalid key
read indexing next record
at end set no-more-records to true
end-read
end-start
.
*><* read next by key of reference paragraph
read-next-by-key.
move corresponding indexing-record to display-record
display display-record end-display
read indexing next record
at end set no-more-records to true
end-read
.
end program indexing.
*><*
*><* Last Update: 20090220
输出：
Reading: 00b02
12345679 0 b++ 2----some 12345679 data here
97654321 0 b++ 2----some 97654321 data here
12349765 0 b++ 2----some 12349765 data here
97651234 0 b++ 2----some 97651234 data here
12345679 0 b++ 2----some 12345679 data here
97654321 0 b++ 2----some 97654321 data here
12349765 0 b++ 2----some 12349765 data here
97651234 0 b++ 2----some 97651234 data here
12345679 0 b++ 2----some 12345679 data here
97654321 0 b++ 2----some 97654321 data here
12349765 0 b++ 2----some 12349765 data here
97651234 0 b++ 2----some 97651234 data here
Seeking >= 00a02
12345679 0 b++ 2----some 12345679 data here
97654321 0 b++ 2----some 97654321 data here
12349765 0 b++ 2----some 12349765 data here
97651234 0 b++ 2----some 97651234 data here
12345679 0 b++ 2----some 12345679 data here
97654321 0 b++ 2----some 97654321 data here
12349765 0 b++ 2----some 12349765 data here
97651234 0 b++ 2----some 97651234 data here
12345679 0 b++ 2----some 12345679 data here
97654321 0 b++ 2----some 97654321 data here
12349765 0 b++ 2----some 12349765 data here
97651234 0 b++ 2----some 97651234 data here
12345689 0 c++13----some 12345689 data here
98654321 0 c++13----some 98654321 data here
12349865 0 c++13----some 12349865 data here
98651234 0 c++13----some 98651234 data here
12345689 0 c++13----some 12345689 data here
98654321 0 c++13----some 98654321 data here
12349865 0 c++13----some 12349865 data here
98651234 0 c++13----some 98651234 data here
12345689 0 c++13----some 12345689 data here
98654321 0 c++13----some 98654321 data here
12349865 0 c++13----some 12349865 data here
98651234 0 c++13----some 98651234 data here
Prime < 87654321
87651234 0 a++ 1----some 87651234 data here
12349865 0 c++13----some 12349865 data here
12349765 0 b++ 2----some 12349765 data here
12348765 0 a++ 1----some 12348765 data here
12345689 0 c++13----some 12345689 data here
12345679 0 b++ 2----some 12345679 data here
12345678 0 a++ 1----some 12345678 data here
在第一次运行时，indexing.dat 不存在。后续运行具有相同的输出：
rewrite key: 12345678
rewrite key: 87654321
rewrite key: 12348765
rewrite key: 87651234
rewrite key: 12345679
rewrite key: 97654321
rewrite key: 12349765
rewrite key: 97651234
rewrite key: 12345689
rewrite key: 98654321
rewrite key: 12349865
rewrite key: 98651234
前置，因为 WRITE INVALID KEY 子句会触发 REWRITE 以允许覆盖密钥和数据。
我知道你的问题很老了。但今天我仍然遇到同样的问题，几个小时后我发现我正在使用的 GnuCOBOL 2.0 和 OpenCobolIDE 4.7.6 在尝试使用 OPEN EXTEND 时会生成错误 35。我以一种不太优雅的方式解决了它，但效果很好。我正在分享我找到的解决方案。
IDENTIFICATION DIVISION.
PROGRAM-ID. CREATFIL.
ENVIRONMENT DIVISION.
CONFIGURATION SECTION.
REPOSITORY.
FUNCTION ALL INTRINSIC.
INPUT-OUTPUT SECTION.
FILE-CONTROL.
SELECT FD-VIRTUAL ASSIGN TO "CADFILE.TXT"
ORGANIZATION IS LINE SEQUENTIAL
ACCESS MODE IS SEQUENTIAL
FILE STATUS IS WS-GET-ERROR.
DATA DIVISION.
FILE SECTION.
FD FD-VIRTUAL.
01 FS-BOOK.
05 NOME PIC X(40).
WORKING-STORAGE SECTION.
77 WS-GET-ERROR PIC XX.
LOCAL-STORAGE SECTION.
01 TB-VIRTUAL.
05 TB-BOOK.
10 LS-NOME PIC X(40).
77 LS-ENTER PIC X.
77 WS-RESP PIC A VALUE "Y".
PROCEDURE DIVISION.
OPEN EXTEND FD-VIRTUAL.
IF WS-GET-ERROR = "35"
OPEN OUTPUT FD-VIRTUAL
END-IF
IF WS-GET-ERROR = "00"
PERFORM UNTIL UPPER-CASE(WS-RESP) NOT = "Y"
DISPLAY X"0D"
DISPLAY "Name: " WITH NO ADVANCING
ACCEPT LS-NOME
WRITE FS-BOOK FROM TB-BOOK
DISPLAY "More record (Y) for YES " WITH NO ADVANCING
DISPLAY "- any key for NOT: "
WITH NO ADVANCING
ACCEPT WS-RESP
IF UPPER-CASE(WS-RESP) NOT = "Y"
EXIT PERFORM
END-IF
END-PERFORM
END-IF.
CLOSE FD-VIRTUAL.
DISPLAY X"0D".
DISPLAY "Press <ENTER> to finish... " WITH NO ADVANCING.
ACCEPT LS-ENTER.
STOP RUN.
END PROGRAM CREATFIL.
遵守指示
OPEN EXTEND FD-VIRTUAL.
IF WS-GET-ERROR = "35"
OPEN OUTPUT FD-VIRTUAL
END-IF
行顺序和索引文件重写与 gnucobol 完美配合。尝试在 termux (android) 上使用 v3.2。
但是当我将我的创作移植到企业 cobol 时，我必须在 jcl 中定义所有分配的名称，并且对于行顺序，我必须使用保存在 uss（z 系统上的 unix）上的文件。也许还有其他解决方案，但这对我有用。
以下行在 jcl 中为您的程序定义 unix 文件名：
//CLCSV1 DD PATH='/z/.../tmpdir/filename'
名称 CLCSV1（不区分大小写）可以分配给文件控制中的文件。当然，选择一个适合您的变量的名称。