Anyone have a data/basic xml parser?
75 views
Skip to first unread message
John R. Stokka
Dec 2, 2025, 3:56:00 PM (2 days ago) Dec 2
to Pick and MultiValue Databases
I'm looking for an xml parser to pick apart an imported data item.
Thought I would see if anyone had such a beast so I don't have to reinvent the wheel.
Thanks in advance. :)
Kevin King
Dec 3, 2025, 8:26:16 AM (yesterday) Dec 3
to mvd...@googlegroups.com
I have some extraction subroutines that use an XPATH-type retrieval.
--
You received this message because you are subscribed to
the "Pick and MultiValue Databases" group.
To post, email to: mvd...@googlegroups.com
To unsubscribe, email to: mvdbms+un...@googlegroups.com
For more options, visit http://groups.google.com/group/mvdbms
---
You received this message because you are subscribed to the Google Groups "Pick and MultiValue Databases" group.
To unsubscribe from this group and stop receiving emails from it, send an email to mvdbms+un...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/mvdbms/3204b971-3011-4a58-8bb0-d717c403cab7n%40googlegroups.com.
John Stokka
Dec 3, 2025, 9:06:57 AM (yesterday) Dec 3
to mvd...@googlegroups.com
To view this discussion visit https://groups.google.com/d/msgid/mvdbms/CALE1pmPYU%3D%2B-E1xV-i%3D%2BOX-NMyDsPcaW0owb83ipicLOWA2dKQ%40mail.gmail.com.
Dave Pomroy
Dec 3, 2025, 10:23:01 AM (yesterday) Dec 3
to mvd...@googlegroups.com
And they work great!
To view this discussion visit https://groups.google.com/d/msgid/mvdbms/CALE1pmPYU%3D%2B-E1xV-i%3D%2BOX-NMyDsPcaW0owb83ipicLOWA2dKQ%40mail.gmail.com.
Kevin King
Dec 3, 2025, 1:36:38 PM (yesterday) Dec 3
to mvd...@googlegroups.com
The functions are specific to U2, but the subroutines are pretty generic, as I recall.
Let me write up some documentation for you and then I'll post.
To view this discussion visit https://groups.google.com/d/msgid/mvdbms/CANhQ6YimzVqgVDBqrVeJfjJyhno1MygjeyiG3j0OaYCQpVW%3DpQ%40mail.gmail.com.
Dick Thiot
Dec 3, 2025, 2:14:00 PM (yesterday) Dec 3
to mvd...@googlegroups.com
It is a lot easier to use a modern language like Python or Javascript to do the work and the coding is easier too. You can use AI to generate the code.
Dick Thiot
RDM Infinity
To view this discussion visit https://groups.google.com/d/msgid/mvdbms/CALE1pmMcV5r%2B%3DYivn%2B8oPM4XA%2BmpDA8viwvdzarnjO1Q1QyMWA%40mail.gmail.com.
John Stokka
Dec 3, 2025, 2:41:34 PM (yesterday) Dec 3
to mvd...@googlegroups.com
Yes, I could do it in c# in my sleep in no time, and find 100 routines on github no doubt. .NET probably has an xml parser built in actually at this point.
But this is a legacy system and we will be doing it old school. ;)
To view this discussion visit https://groups.google.com/d/msgid/mvdbms/CAAbj8ioOeAB2z%3DknryNtpQTZKPVVG25CSCDt1Uk5eyHJc_WJgQ%40mail.gmail.com.
Tom Marracci
Dec 3, 2025, 2:47:50 PM (yesterday) Dec 3
to mvd...@googlegroups.com
For what it's worth, here's what I use:
DECODE.XML
001 SUBROUTINE (JSON,XML)
002 *
003 * DECODE AN XML DOCUMENT INTO A JSON STRUCTURE
004 *
005 * USE STATE MACHINE TO TRAVERSE THE XML TEXT
006 * USE CHUNKS AND ARRAYS TO COMBAT PICK STRING HANDLING
007 *
008 *
009 *
010 * DECODE XML STATE MACHINE AND CONSTANTS
011 *
012 * USE STATE MACHINE TO TRAVERSE THE XML TEXT
013 * USE CHUNKS AND ARRAYS TO COMBAT PICK STRING HANDLING
014 *
015 * CONSIDER CHANGE FOR STRINGS:
016 * MAKE LIST OF SYMBOLS TO REPLACE
017 * MARK BEGINNING OF STRING
018 * EXTRACT STRING AT THE END
019 * AND CHANGE ENCODED SYMBOLS
020 TFLG=INDEX(SYSTEM(15),'T',1)
021 * CHARACTER CLASSES
022 EQU C_ALPHA TO 1
023 EQU C_DIGIT TO 2
024 EQU C_UNDER TO 3
025 EQU C_SLASH TO 4
026 EQU C_OPEN TO 5
027 EQU C_CLOSE TO 6
028 EQU C_BANG TO 7
029 EQU C_WHITE TO 8
030 EQU C_DQUOTE TO 9
031 EQU C_HASH TO 10
032 EQU C_AMP TO 11
033 EQU C_SQUOTE TO 12
034 EQU C_COLON TO 13
035 EQU C_SEMICOLON TO 14
036 EQU C_EQUAL TO 15
037 EQU C_QUES TO 16
038 EQU C_LBR TO 17
039 EQU C_RBR TO 18
040 EQU C_ESC TO 19
041 EQU C_DASH TO 20
042 EQU C_ETC TO 21
043 EQU C_LENGTH TO 21
044 DIM CHARS(128)
045 MATPARSE CHARS FROM '0,0,0,0,0,0,0,0,0,8,8,0,0,8,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,8,7,9,10,21,21,11,12,21,21,21,21,21,20,21,4,2,2,2,2,2,2,2,2,2,2,13,14,5,15,6,16,0,1,1,1,1,1,1,1,1,1
,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,17,19,18,21,21,21,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,21,21,21,21' USING ','
046 CHARS(SEQ('_')+1) = C_ALPHA
047 *
048 * STATES
049 EQU S_START TO 1
050 EQU S_ENTER TO 2
051 EQU S_CMT TO 3
052 EQU S_CMT2 TO 4
053 EQU S_CMT3 TO 5
054 EQU S_CMT4 TO 6
055 EQU S_CMT5 TO 7
056 EQU S_PREMB TO 8
057 EQU S_PREMB1 TO 9
058 EQU S_KEY TO 10
059 EQU S_BOOL1 TO 11
060 EQU S_BOOL2 TO 12
061 EQU S_VALUE TO 13
062 EQU S_END1 TO 14
063 EQU S_ENDKEY TO 15
064 EQU S_SYM1 TO 16
065 EQU S_SYM2 TO 17
066 EQU S_SYM3 TO 18
067 EQU S_ATTR TO 19
068 EQU S_PROCBOOL TO -2
069 EQU S_PROCVALU TO -3
070 EQU S_PROCSYM TO -4
071 EQU S_LENGTH TO 19
072 EQU S_DSTRING TO 20; * DOUBLE QUOTED STRING
073 EQU S_SSTRING TO 21; * SINGLE QUOTED STRING
074 STATE.TXT = 'START,ENTER,CMT,CMT2,CMT3,CMT4,CMT5,PREM,PREMB1,KEY,BOOL1,BOOL2,VALUE,END1,ENDKEY,SYM1,SYM2,SYM3,ATTR'
075 *
076 * TRANSITIONS
077 * A-Z 0-9 _ / < > ! <SP> " # & ' : ; = ? [ ] \ - ETC
078 *START ENT START
079 *ENTER KEY CMT ENTER PREMB
080 *CMT CMT2
081 *CMT2 CMT3
082 *CMT3 CMT4
083 *CMT4 CMT5
084 *CMT5 START
085 *PREMB PREMB1
086 *PREMB1 START
087 *KEY KEY KEY KEY VALUE BOOL1 KEY
088 *BOOL1 BOOL2 VALUE
089 *BOOL2 PROCBOOL
090 *VALUE END1 SYM1
091 *END1 ENDKEY
092 *ENDKEY PROCVALU
093 *SYM1 SYM2 SYM3
094 *SYM2 SYM2 PROCSYM
095 *SYM3 SYM3 SYM3 PROCSYM
096 *ATTR BOOL2 VALU
097 *
098 DIM STATES(S_LENGTH,C_LENGTH)
099 IF TFLG AND 0 THEN
100 PRINT 'BUILDING STATE ARRAY'
101 MAT STATES = 0
102 * SETUP STATES THAT LOOP UNTIL EXIT CHARACTER IS FOUND
103 FOR I = 1 TO C_ETC
104 STATES(S_CMT3,I) = S_CMT3
105 STATES(S_CMT4,I) = S_CMT3
106 STATES(S_CMT5,I) = S_CMT3
107 STATES(S_PREMB,I) = S_PREMB
108 STATES(S_VALUE,I) = S_VALUE
109 STATES(S_ENDKEY,I) = S_ENDKEY
110 STATES(S_ATTR,I) = S_ATTR
111 STATES(S_BOOL1,I) = S_ATTR
112 NEXT I
113 STATES(S_START,C_OPEN) = S_ENTER
114 STATES(S_START,C_WHITE) = S_START
115 STATES(S_ENTER,C_ALPHA) = S_KEY
116 STATES(S_ENTER,C_UNDER) = S_KEY
117 STATES(S_ENTER,C_BANG) = S_CMT
118 STATES(S_ENTER,C_WHITE) = S_ENTER
119 STATES(S_ENTER,C_QUES) = S_PREMB
120 STATES(S_ENTER,C_SLASH) = S_ENDKEY
121 STATES(S_CMT,C_DASH) = S_CMT2
122 STATES(S_CMT2,C_DASH) = S_CMT3
123 STATES(S_CMT3,C_DASH) = S_CMT4; * SPECIAL HANDLING HERE - ANYTHING ELSE IS SKIPPED
124 STATES(S_CMT4,C_DASH) = S_CMT5
125 STATES(S_CMT5,C_CLOSE) = S_START
126 STATES(S_PREMB,C_QUES) = S_PREMB1; * ANYTHING ELSE IS SKIPPED
127 STATES(S_PREMB1,C_CLOSE) = S_START
128 STATES(S_KEY,C_ALPHA) = S_KEY
129 STATES(S_KEY,C_DIGIT) = S_KEY
130 STATES(S_KEY,C_UNDER) = S_KEY
131 STATES(S_KEY,C_COLON) = S_KEY
132 STATES(S_KEY,C_CLOSE) = S_VALUE
133 STATES(S_KEY,C_WHITE) = S_BOOL1
134 STATES(S_BOOL1,C_WHITE) = S_BOOL1
135 STATES(S_BOOL1,C_SLASH) = S_BOOL2
136 STATES(S_BOOL1,C_CLOSE) = S_VALUE
137 STATES(S_BOOL2,C_CLOSE) = S_PROCBOOL
138 STATES(S_VALUE,C_OPEN) = S_END1
139 STATES(S_END1,C_SLASH) = S_ENDKEY
140 STATES(S_END1,C_ALPHA) = S_KEY
141 STATES(S_END1,C_UNDER) = S_KEY
142 STATES(S_ENDKEY,C_CLOSE) = S_PROCVALU
143 STATES(S_VALUE,C_AMP) = S_SYM1
144 STATES(S_SYM1,C_ALPHA) = S_SYM2
145 STATES(S_SYM1,C_HASH) = S_SYM3
146 STATES(S_SYM2,C_ALPHA) = S_SYM2
147 STATES(S_SYM2,C_SEMICOLON) = S_PROCSYM
148 STATES(S_SYM3,C_ALPHA) = S_SYM3
149 STATES(S_SYM3,C_DIGIT) = S_SYM3
150 STATES(S_SYM3,C_SEMICOLON) = S_PROCSYM
151 STATES(S_ATTR,C_SLASH) = S_BOOL2
152 STATES(S_ATTR,C_CLOSE) = S_VALUE
153 IF TFLG THEN
154 MATBUILD TMP FROM STATES USING ','
155 CALL FORMS.STREAM(1)
156 PRINT TMP
157 CALL FORMS.STREAM(0)
158 END
159 END ELSE
160 MATPARSE STATES FROM '0,0,0,0,2,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,10,0,10,15,0,0,3,2,0,0,0,0,0,0,0,8,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,4,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
,0,0,0,5,0,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,6,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,7,5,5,5,5,5,5,1,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,9,8,8,8,8,8,0,0,0,0
,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,10,10,10,0,0,13,0,11,0,0,0,0,10,0,0,0,0,0,0,0,0,19,19,19,12,19,13,19,11,19,19,19,19,19,19,19,19,19,19,19,19,19,0,0,0,0,0,-2,0,0,0,0,0,0,0,0,0,0,0,0,0,0
,0,13,13,13,13,14,13,13,13,13,13,16,13,13,13,13,13,13,13,13,13,13,10,0,10,15,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,15,15,15,15,15,-3,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,17,0,0,0,0,0
,0,0,0,18,0,0,0,0,0,0,0,0,0,0,0,17,0,0,0,0,0,0,0,0,0,0,0,0,-4,0,0,0,0,0,0,0,18,18,0,0,0,0,0,0,0,0,0,0,0,-4,0,0,0,0,0,0,0,19,19,19,12,19,13,19,19,19,19,19,19,19,19,19,19,19,19,19,19,19' U
SING ','
161 END
162 * MODES
163 *
164 * TYPES (FROM PHP)
165 * AND USED IN RESULT TO INDICATE TYPE OF EACH ELEMENT
166 EQU IS_NULL TO 0
167 EQU IS_LONG TO 1
168 EQU IS_DOUBLE TO 2
169 EQU IS_BOOL TO 3
170 EQU IS_ARRAY TO 4
171 EQU IS_OBJECT TO 5
172 EQU IS_STRING TO 6
173 *
174 * SPECIAL SYMBOLS
175 *
176 DIM SYMBOLS(4)
177 SYMBOLS(1) = 'amp]lt]gt]quot'
178 SYMBOLS(2) = '&]<]>]"'
179 SYMBOLS(3) = 'a]am]amp]l]lt]g]gt]q]qu]quo]quot'
180 SYMBOLS(4) = '0123456789ABCDEFabcdef'
181 *
182 *
183 EQU CHUNK.SIZE TO 512
184 *
185 * INITIALIZE RESULTING OBJECT IN RBUF
186 RSIZE = 1000
187 DIM RBUF(RSIZE)
188 RBUF = ''
189 RP = 1
190 RBUF(1) = IS_OBJECT
191 *
192 * INITIALIZE STRING BUFFER TO 16 SEGMENTS
193 SSIZE = 16
194 DIM SBUF(SSIZE)
195 *
196 * INITIALIZE RUNTIME PARAMETERS
197 STACK = ''
198 XLEN = LEN(XML)
199 STATE = S_START
200 KEY = ''
201 GOSUB RESET.VAL
202 SYM = ''
203 ATTR = ''
204 JSON = ''
205 STR.STATE = ''
206 * PROCESS IN CHUNKS TO COMBAT PICK STRING HANDLING
207 FOR I = 1 TO XLEN STEP CHUNK.SIZE
208 IXML = XML[I,CHUNK.SIZE]
209 XXLEN = LEN(IXML)
210 FOR J = 1 TO XXLEN
211 CH = IXML[J,1]
212 IF CH = CHAR(254) THEN CH = CHAR(10)
213 CHVAL = SEQ(CH)
214 IF CHVAL > 128 THEN CHCLS = C_ETC ELSE CHCLS = CHARS(CHVAL+1)
215 IF CHCLS = 0 THEN CHCLS = C_ETC
216 IF STATE = S_ATTR AND CHCLS = C_DQUOTE THEN
217 STR.STATE = STATE
218 STATE = S_DSTRING
219 CONTINUE
220 END ELSE IF STATE = S_ATTR AND CHCLS = C_SQUOTE THEN
221 STR.STATE = STATE
222 STATE = S_SSTRING
223 CONTINUE
224 END ELSE IF STATE = S_DSTRING THEN
225 IF CHCLS = C_DQUOTE THEN STATE = STR.STATE
226 CONTINUE
227 END ELSE IF STATE = S_SSTRING THEN
228 IF CHCLS = C_SQUOTE THEN STATE = STR.STATE
229 CONTINUE
230 END ELSE
231 NEXT.STATE = STATES(STATE,CHCLS)
232 END
233 IF NEXT.STATE > 0 THEN
234 ON NEXT.STATE GOSUB DO.START,DO.ENTER,DO.CMT,DO.CMT2,DO.CMT3,DO.CMT4,DO.CMT5,DO.PREMB,DO.PREMB1,DO.KEY,DO.BOOL1,DO.BOOL2,DO.VALUE,DO.END1,DO.ENDKEY,DO.SYM1,DO.SYM2,DO.SYM3,DO.ATT
R
235 STATE = NEXT.STATE
236 END ELSE IF NEXT.STATE < 0 THEN
237 ON ABS(NEXT.STATE)-1 GOSUB DO.PROCBOOL,DO.PROCVALU,DO.PROCSYM
238 END ELSE
239 DO.ERROR:
240 PRINT 'ERROR'
241 PRINT '@ ':I:':':J
242 PRINT CH,CHVAL,CHCLS
243 PRINT STATE:' ':FIELD(STATE.TXT,',',STATE)
244 PRINT IXML[J,32]
245 JSON=''
246 RETURN
247 END
248 NEXT J
249 NEXT I
250 IF RP THEN MATBUILD JSON FROM RBUF,1,RP
251 *
252 RETURN
253 DO.START:
254 RETURN
255 * START A NEW KEY
256 DO.ENTER:
257 KEY = ''
258 ATTR = ''
259 GOSUB RESET.VAL
260 RETURN
261 *
262 * SKIP OVER A COMMENT
263 DO.CMT:
264 RETURN
265 DO.CMT2:
266 RETURN
267 DO.CMT3:
268 RETURN
269 DO.CMT4:
270 RETURN
271 DO.CMT5:
272 RETURN
273 *
274 * SKIP OVER THE XML PREAMBLE
275 DO.PREMB:
276 RETURN
277 DO.PREMB1:
278 RETURN
279 *
280 * ACCUMULATE A KEY
281 DO.KEY:
282 IF CH = ':' THEN
283 KEY = ''; * DISCARD NAMESPACE
284 END ELSE
285 KEY := CH
286 END
287 RETURN
288 *
289 * PROCESS A POSSIBLE BOOLEAN VALUE
290 DO.BOOL1:
291 RETURN
292 DO.BOOL2:
293 RETURN
294 *
295 * KEY OBTAINED
296 * UPDATE THE PATH
297 * AND ACCUMULATE A VALUE
298 DO.VALUE:
299 IF CHCLS = C_CLOSE THEN
300 * CHECK TOP OF THE STACK
301 * IF NOT AN OBJECT, MAKE IT AN OBJECT NOW
302 IF STACK='' THEN M = 1 ELSE M = STACK<1>
303 IF RBUF(M)<1,1> = IS_ARRAY THEN
304 GOSUB ADD.NODE
305 RBUF(M)<1,2,-1> = N
306 M = N
307 GOSUB ADD.NODE
308 RBUF(M) = IS_OBJECT
309 RBUF(M)<1,2> = KEY
310 RBUF(M)<1,3> = N
311 END ELSE
312 IF RBUF(M)<1,1> # IS_OBJECT THEN RBUF(M) = IS_OBJECT
313 GOSUB ADD.NODE
314 GOSUB ADD.KEY
315 END
316 INS N BEFORE STACK<1>
317 GOSUB RESET.VAL
318 KEY = ''
319 ATTR = ''
320 END ELSE
321 GOSUB APPEND.CHAR
322 END
323 RETURN
324 *
325 *
326 * PROCESS END OF A KEY
327 DO.END1:
328 RETURN
329 DO.ENDKEY:
330 IF CHCLS = C_SLASH THEN
331 KEY = ''
332 ATTR = ''
333 END ELSE IF CH = ':' THEN
334 KEY = ''; * DISCARD NAMESPACE
335 END ELSE
336 KEY := CH
337 END
338 RETURN
339 *
340 * PROCESS A KEY WITH BOOLEAN VALUE
341 * OF THE FORM <NAME />
342 * SET NAME TO BE A BOOLEAN WITH TRUE VALUE
343 DO.PROCBOOL:
344 M = STACK<1>
345 IF RBUF(M)<1,1> # IS_OBJECT THEN RBUF(M) = IS_OBJECT
346 GOSUB ADD.NODE
347 GOSUB ADD.KEY
348 RBUF(N) = IS_BOOL
349 RBUF(N)<1,2> = 1
350 KEY = ''
351 ATTR = ''
352 STATE = S_START
353 RETURN
354 *
355 *
356 * PROCESS A SCALAR VALUE
357 * DESTINATION WAS PLACED ONTO STACK BY VALUE
358 * IF IT'S STILL NULL, THEN IT IS A SCALAR THAT WE NEED TO UPDATE
359 * DETERMINE THE TYPE AND UPDATE THE VALUE
360 DO.PROCVALU:
361 * KEY COMES FROM ENDKEY
362 * IT SHOULD MATCH THE KEY BEING UPDATED BUT WE'RE NOT CHECKING
363 * VAL WAS BUILT UP
364 * TOP OF STACK IS WHERE THIS VALUE GOES
365 * DETERMININE IT'S TYPE
366 * AND UPDATE THE RESULT
367 N = STACK<1>
368 DEL STACK<1>
369 GOSUB GET.VAL
370 IF RBUF(N) = IS_NULL THEN
371 RBUF(N) = IS_STRING
372 IF VAL MATCHES '1N0N' OR VAL MATCHES '-1N0N' THEN
373 RBUF(N) = IS_LONG
374 END ELSE IF VAL MATCHES '1N0N.1N0N' OR VAL MATCHES '-1N0N.1N0N' THEN
375 RBUF(N) = IS_DOUBLE
376 END
377 RBUF(N)<1,2> = VAL
378 END
379 KEY = ''
380 ATTR = ''
381 GOSUB RESET.VAL
382 STATE = S_START
383 RETURN
384 *
385 * HTML ENTITY ENCOUNTERED
386 * PARSE AND CHANGE IT TO THE DESIRED CHARACTER
387 DO.SYM1:
388 SYM = ''
389 RETURN
390 DO.SYM2:
391 SYM := CH
392 LOCATE SYM IN SYMBOLS(3)<1> SETTING P ELSE RETURN TO DO.ERROR
393 RETURN
394 DO.SYM3:
395 IF CH = '#' THEN
396 SYM = CH
397 END ELSE IF INDEX(SYMBOLS(4),CH,1) THEN
398 SYM := CH
399 END ELSE
400 RETURN TO DO.ERROR
401 END
402 RETURN
403 *
404 * COMPLETE TRANSLATION OF HTML ENTITIES
405 * ONLY CERTAIN ENTITIES ARE SUPPORTED
406 * SEE THE SYMBOL ARRAY FOR DETAILS
407 DO.PROCSYM:
408 * DEDUCE SYMBOL
409 IF SYM[1,1]='#' THEN
410 SYM = CHAR(XTD(SYM[2,LEN(SYM)]))
411 END ELSE
412 LOCATE SYM IN SYMBOLS(1)<1> SETTING P ELSE RETURN TO DO.ERROR
413 SYM = SYMBOLS(2)<1,P>
414 END
415 CH = SYM
416 GOSUB APPEND.CHAR
417 SYM = ''
418 STATE = S_VALUE
419 RETURN
420 *
421 * ACCUMULATE A TAGS ATTRIBUTES
422 * JSON HAS NO NEED FOR THESE SO THEY WILL BE DISCARDED
423 DO.ATTR:
424 ATTR := CH
425 RETURN
426 *
427 *
428 * HANDLE QUOTED STRINGS
429 DO.DSTRING:
430 DO.SSTRING:
431 *
432 * APPEND CHAR TO VAL
433 *
434 APPEND.CHAR:
435 IF LEN(SBUF(SP)) GE CHUNK.SIZE THEN
436 SP += 1
437 IF SP GE SSIZE THEN
438 SSIZE *= 2
439 TMP = SBUF
440 DIM SBUF(SSIZE)
441 SBUF = TMP
442 END
443 END
444 SBUF(SP) := CH
445 RETURN
446 *
447 RESET.VAL:
448 VAL = ''
449 SP = 1
450 SBUF = ''
451 RETURN
452 *
453 GET.VAL:
454 MATBUILD VAL FROM SBUF
455 CONVERT CHAR(254) TO '' IN VAL
456 RETURN
457 *
458 * ADD A NEW NODE TO THE RESULT
459 * RESIZE AS NEEDED
460 * RETURNS N - POSITION OF THE NEW NODE
461 ADD.NODE:
462 IF RP+1 > RSIZE THEN
463 RSIZE *= 2
464 DIM RBUF(RSIZE)
465 END
466 RP += 1
467 N = RP
468 RBUF(N) = IS_NULL
469 RETURN
470 *
471 * ADD KEY AS A NODE OR ARRAY
472 * M IS PARENT, N IS NODE
473 * IF ARRAY, UPDATE ARRAY AND IF NECESSARY GET ANOTHER NODE
474 ADD.KEY:
475 LOCATE KEY IN RBUF(M)<1,2> SETTING PV THEN
476 RBUF(N) = RBUF(M)
477 RBUF(M) = IS_ARRAY
478 RBUF(M)<1,2> = N
479 GOSUB ADD.NODE
480 RBUF(N) = IS_OBJECT
481 RBUF(N)<1,2> = KEY
482 RBUF(M)<1,2,2> = N
483 M = N
484 GOSUB ADD.NODE
485 RBUF(M)<1,3> = N
486 END ELSE
487 RBUF(M)<1,2,-1> = KEY
488 RBUF(M)<1,3,-1> = N
489 END
490 RETURN
491 DUMP:
492 FOR DI = 1 TO RP
493 PRINT DI'R%3 ':RBUF(DI)
494 NEXT DI
495 PRINT STR('-',10)
496 RETURN
497 END
This creates a dynamic array with named keys. Then I have a subroutine that I use to access individual elements of the structure this decode creates:
GET.JSON
L9999
001 SUBROUTINE (JSON,KEY,VAL)
002 *
003 * GET AN ELEMENT, ARRAY, OR OBJECT FROM A JSON OBJECT
004 *
005 * IF JSON IS {} THEN KEY IS STRING (WITH MULTIPLE KEYS SEPARATED WITH DOT (.))
006 * IF AN ELEMENT IS AN ARRAY, USE THE INDEX INSTEAD OF A NAME
007 * FOR EXAMPLE
008 * {carriers:[{id:"x"},{id:"y"},{id:"z"}]}
009 * COULD USE carriers.0.id TO GET THE ID OF FIRST ELEMENT
010 * OR ARRAY NOTATION carriers[0].id
011 * IF JSON IS [] THEN KEY IS AN INDEX STARTING AT 0
012 *
013 * DEPENDING ON THE SELECTION, VAL CAN BE A SCALAR, ARRAY, OR OBJECT
014 * IF ARRAY OR OBJECT, THE RESULT IS MODIFIED TO INDEX CORRECTLY TO ITSELF
015 * THUS IT IS ANOTHER JSON OBJECT
016 *
017 * TYPES (FROM PHP)
018 * AND USED IN RESULT TO INDICATE TYPE OF EACH ELEMENT
019 EQU IS_NULL TO 0
020 EQU IS_LONG TO 1
021 EQU IS_DOUBLE TO 2
022 EQU IS_BOOL TO 3
023 EQU IS_ARRAY TO 4
024 EQU IS_OBJECT TO 5
025 EQU IS_STRING TO 6
026 *
027 * FIND THE POSITION OF THE KEY PATH
028 * IF ARRAY, SHOULD BE A NUMERIC ZERO BASED INDEX
029 * SEPERATE COMPONENTS WITH DOTS (.)
030 * IF NOT FOUND, RETURN NULL
031 VAL = ''
032 *
033 * TO SPEED IT UP, MAKE A COPY AS A DIMENSIONED ARRAY
034 M = DCOUNT(JSON,CHAR(254))
035 IF M < 1 THEN RETURN
036 DIM JSONP(M); JSONP = JSON
037 TYPE = JSONP(1)<1,1>
038 N = 1
039 IF TYPE=IS_ARRAY THEN
040 IF KEY MATCHES '1N0N' THEN N = JSONP(N)<1,2,KEY+1> ELSE RETURN
041 END ELSE IF TYPE=IS_OBJECT THEN
042 CONVERT '[]' TO '.' IN KEY; * CONVERT SHORT HAND ARRAY INDEXING TO DOTS
043 * AND REMOVE CLOSING BRACKET (])
044 FOR I = 1 TO DCOUNT(KEY,'.')
045 K = FIELD(KEY,'.',I)
046 IF JSONP(N)<1,1> = IS_ARRAY THEN
047 IF K MATCHES '1N0N' THEN
048 N = JSONP(N)<1,2,K+1>; * ZERO INDEXED
049 IF N='' THEN RETURN; * OUT OF RANGE
050 END ELSE
051 RETURN
052 END
053 END ELSE IF JSONP(N)<1,1> = IS_OBJECT THEN
054 LOCATE K IN JSONP(N)<1,2> SETTING Q THEN N = JSONP(N)<1,3,Q> ELSE RETURN
055 END ELSE
056 RETURN
057 END
058 NEXT I
059 END
060 *
061 *
062 * WE FOUND IT. DETERMINE WHAT IT IS AND THE VALUE TO RETURN
063 * FOR ARRAY AND OBJECT, DEREFERENCE THE COMPONENTS
064 * AND REINDEX RELEVANT TO THE RESULT
065 IF N THEN
066 TYPE = JSONP(N)<1,1>
067 BEGIN CASE
068 CASE TYPE=IS_NULL; VAL = ''
069 CASE TYPE=IS_LONG; VAL = JSONP(N)<1,2>
070 CASE TYPE=IS_DOUBLE; VAL = JSONP(N)<1,2>
071 CASE TYPE=IS_BOOL; VAL = JSONP(N)<1,2>
072 CASE TYPE=IS_STRING; VAL = JSONP(N)<1,2>
073 CASE TYPE=IS_ARRAY OR TYPE=IS_OBJECT
074 * ARRAY AND OBJECT CONTAIN ATTRIBUTES INTO THE RECORD
075 * ADJUST THESE TO BE RELEVANT TO THE RESULT
076 * SOME REORDERING MAY TAKE PLACE BUT THAT'S OK
077 * AS LONG AS WE USE THE API TO ACCESS IT
078 VAL = JSONP(N)
079 P = 1; * LAST ELEMENT IN RESULT
080 N = 1; * CURRENT WORKING ITEM
081 LOOP
082 TYPE = VAL<N,1>
083 IF TYPE=IS_ARRAY THEN
084 FOR I = 1 TO DCOUNT(VAL<N,2>,CHAR(252))
085 M = VAL<N,2,I>; * THE OLD INDEX
086 P += 1
087 VAL<P> = JSONP(M)
088 VAL<N,2,I> = P
089 NEXT I
090 END ELSE IF TYPE=IS_OBJECT THEN
091 FOR I = 1 TO DCOUNT(VAL<N,3>,CHAR(252))
092 M = VAL<N,3,I>; * OLD INDEX
093 P += 1
094 VAL<P> = JSONP(M)
095 VAL<N,3,I> = P
096 NEXT I
097 END
098 WHILE VAL<N+1>#'' DO
099 N += 1
100 REPEAT
101 END CASE
102 END
103 RETURN
104 END
Tom
From: mvd...@googlegroups.com <mvd...@googlegroups.com> on behalf of John Stokka <jrst...@gmail.com>
Sent: Wednesday, December 3, 2025 11:40 AM
To: mvd...@googlegroups.com <mvd...@googlegroups.com>
Subject: Re: [mvdbms] Anyone have a data/basic xml parser?
Sent: Wednesday, December 3, 2025 11:40 AM
To: mvd...@googlegroups.com <mvd...@googlegroups.com>
Subject: Re: [mvdbms] Anyone have a data/basic xml parser?
To view this discussion visit
https://groups.google.com/d/msgid/mvdbms/CANhQ6Yh0p5Nf0E4-_mXOrPtUpz7qC2dSv-Y3DPnMeqn2Uz7aDQ%40mail.gmail.com.
Kevin King
Dec 3, 2025, 3:04:50 PM (yesterday) Dec 3
to mvd...@googlegroups.com
Hopefully attachments are allowable. While this code uses a bit of U2 syntax, it requires no platform-specific functionality.
Kevin King
Dec 3, 2025, 3:06:50 PM (yesterday) Dec 3
to mvd...@googlegroups.com
I should also note that my routines leave things in XML until they can be converted to a name/value format. This works around U2's limitation on the depth limitation of attributes, values, and subvalues.
John Stokka
Dec 3, 2025, 3:09:58 PM (yesterday) Dec 3
to mvd...@googlegroups.com
Tom,
That's kinda handy, converting the XML to JSON first then you have a json parser... Cool. Thanks!
To view this discussion visit https://groups.google.com/d/msgid/mvdbms/CH3PR19MB794689C710627238C3F90028DDD9A%40CH3PR19MB7946.namprd19.prod.outlook.com.
John Stokka
Dec 3, 2025, 3:10:29 PM (yesterday) Dec 3
to mvd...@googlegroups.com
Kevin,
That's exactly what I'm looking for! Thanks so much!
On Wed, Dec 3, 2025 at 2:04 PM Kevin King <ke...@precisonline.com> wrote:
Hopefully attachments are allowable. While this code uses a bit of U2 syntax, it requires no platform-specific functionality.
--
You received this message because you are subscribed to
the "Pick and MultiValue Databases" group.
To post, email to: mvd...@googlegroups.com
To unsubscribe, email to: mvdbms+un...@googlegroups.com
For more options, visit http://groups.google.com/group/mvdbms
---
You received this message because you are subscribed to the Google Groups "Pick and MultiValue Databases" group.
To unsubscribe from this group and stop receiving emails from it, send an email to mvdbms+un...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/mvdbms/CALE1pmNg-ZFpbv8QeLU_fPw2Sfs9GWsmfywu42pfj4tgkHC2aA%40mail.gmail.com.
Bob Dubery
2:50 AM (14 hours ago) 2:50 AM
to Pick and MultiValue Databases
I tried that with ChatGPT. It was very good at giving me a program that would extract from a piece of XML with a known layout (which I'd already figured out). It's harder (for me and for ChatGPT) to write a generalised tool that just unpacks a piece of XML and presents it to another U2 application.
There are two many programs of the first type at my place of work, I wanted a more generalised and thus reusable solution (though really all you do is move the hard coded identifying of tags to another program).
I don't feel I can share the code here because it's the IP of the company I currently work for, but it was not difficult to write a program that would present the XML as a "flat" dynamic array. EG 1 attribute per tag: 1st value is tag name, 2nd value is content (if any), 3rd value is sub-valued attribute names, 4th is values per attribute and so on. You lose the structure of the XML (this tag has that as a parent, has these as children) but that turned out to not be much of a problem for my application.
Some years ago I spent a couple of weeks writing a suite of subroutines that gave DOM like functions for reading or generating XML. They had names like GetNextChild, AddNextChild and so on. And maybe this is the trick. Think of the functions you need in plain language and give them good names, then figure out a data structure that will allow all those things to happen, AND THEN write a subroutine per function.
All of what I describe above was done in UniVerse BASIC, or on MVOn, with no resorting to other languages. I don't have the system privileges these days to install other languages or libraries. When I did the DOM-like functions I had root access, I had PERL available, I could install libraries if I wanted, but it was still hard to get back something that fitted well with UniVerse (I was keen that other UV programmers should be able to easily use these tools) and that's why in the end I did it all with what a not terribly recent version of UniVerse gave me.
So it can be done. The trick is more mental than about programming chops.
There are two many programs of the first type at my place of work, I wanted a more generalised and thus reusable solution (though really all you do is move the hard coded identifying of tags to another program).
I don't feel I can share the code here because it's the IP of the company I currently work for, but it was not difficult to write a program that would present the XML as a "flat" dynamic array. EG 1 attribute per tag: 1st value is tag name, 2nd value is content (if any), 3rd value is sub-valued attribute names, 4th is values per attribute and so on. You lose the structure of the XML (this tag has that as a parent, has these as children) but that turned out to not be much of a problem for my application.
Some years ago I spent a couple of weeks writing a suite of subroutines that gave DOM like functions for reading or generating XML. They had names like GetNextChild, AddNextChild and so on. And maybe this is the trick. Think of the functions you need in plain language and give them good names, then figure out a data structure that will allow all those things to happen, AND THEN write a subroutine per function.
All of what I describe above was done in UniVerse BASIC, or on MVOn, with no resorting to other languages. I don't have the system privileges these days to install other languages or libraries. When I did the DOM-like functions I had root access, I had PERL available, I could install libraries if I wanted, but it was still hard to get back something that fitted well with UniVerse (I was keen that other UV programmers should be able to easily use these tools) and that's why in the end I did it all with what a not terribly recent version of UniVerse gave me.
So it can be done. The trick is more mental than about programming chops.
Reply all
Reply to author
Forward
0 new messages