Creating a Random File

[Dimster]

I've been trying to create a Random File for 50 Events. Each event has 9 aspects to track. I'm not sure where I'm going wrong. In the following coding I am trying to write to Event #5 and Event #3 only. I'm then trying to retrieve and print to screen only some of the data from each of Event 5 and Event 3. What appears on the screen isn't what I thought was put into the records. Not sure exactly where I'm screwing up here.

  [ You are not allowed to view this attachment ]  
 

[SMcNeill]

Code: QB64: [Select]

1. TYPE EventN
2.     LastEvent AS SINGLE '4
3.     TotEvent AS SINGLE '8
4.     EventW AS SINGLE '12
5.     EventV AS SINGLE '16
6.     EventRP AS SINGLE '20
7.     EventRPW AS SINGLE '24
8.     EventRPV AS SINGLE '28
9.     EventDec AS STRING * 7 '35
10.     EventExp AS STRING * 15 '50
11. END TYPE
12.  
13. DIM SHARED EvRecord AS EventN
14.  
15. OPEN "EventRecords" FOR RANDOM AS #700 LEN = 50
16.  
17. EvRecord.LastEvent = 5 'Fill the UDT with all the data fields at once.
18. EvRecord.TotEvent = 5
19. EvRecord.EventW = 5
20. EvRecord.EventV = 5
21. EvRecord.EventRP = 5
22. EvRecord.EventRPW = 5
23. EvRecord.EventRPV = 5
24. EvRecord.EventDec = "None"
25. EvRecord.EventExp = "Value de minimis"
26. PUT #700, 5, EvRecord 'Put the whole record into position #5
27.  
28. EvRecord.LastEvent = 3 'Fill the whole UDT with all the data fields at once.
29. EvRecord.TotEvent = 3
30. EvRecord.EventW = 3
31. EvRecord.EventV = 3
32. EvRecord.EventRP = 3
33. EvRecord.EventRPW = 3
34. EvRecord.EventRPV = 3
35. EvRecord.EventDec = "Proceed"
36. EvRecord.EventExp = "Confidence 80%"
37. PUT #700, 3, EvRecord 'Put the whole record into position #3
38. CLOSE 700
39.  
40. OPEN "EventRecords" FOR RANDOM AS #700 LEN = 50
41. GET #700, 5, EvRecord
42. LOCATE 5, 10: PRINT EvRecord.LastEvent
43. LOCATE 10, 10: PRINT EvRecord.EventDec
44. LOCATE 10, 30: PRINT EvRecord.EventExp
45.  
46. GET #700, 3, EvRecord
47. LOCATE 15, 10: PRINT EvRecord.LastEvent
48. LOCATE 20, 10: PRINT EvRecord.EventDec
49. LOCATE 20, 30: PRINT EvRecord.EventExp
50. CLOSE 700

[Dimster]

Thanks Steve - in your version of the Type statement there are remarked values of 4,8,12 ...50. What do these values mean.

I didn't realize 1 Put statement writes the entire record. Thanks again for the help.

[Dimster]

I do see where each single is 4 bytes but the strings at '*7' and '*15' are a little confusing. The final remark = 50. I gather that's no coincidence that the Random File is to have 50 records. Is 50 records the same thing as 50 bytes?

OH, the light has struck - it's not numbers of records the LEN is calling for but the Length of EACH record.

Sometimes Steve I wonder why I read it over and over again and still don't see the forest for the trees.

Appreciate your patience

[SMcNeill]

I do see where each single is 4 bytes but the strings at '*7' and '*15' are a little confusing. The final remark = 50. I gather that's no coincidence that the Random File is to have 50 records. Is 50 records the same thing as 50 bytes?

OH, the light has struck - it's not numbers of records the LEN is calling for but the Length of EACH record.

Sometimes Steve I wonder why I read it over and over again and still don't see the forest for the trees.

Appreciate your patience

Anytime.  :)

You’re correct, the LEN is looking for the length of the records — which is why you need to have fixed length strings for your TYPE.  When looking at your previous code, it seemed as if you were expecting a 50 byte record, so I tried to decipher what was needed for your example records 3 and 5, so I guessed at the size 7 and 15 for the string’s limit. 

Usually, instead of hard coding a value, it’s often better to just set the record length with the LEN statement.

Open “my file.rnd” FOR RANDOM AS #1 LEN = LEN(ReferenceVariable)

(In this case, it’d be LEN = LEN(EvRecord).)

The main advantage to do it like so?

1) You don’t have to manually count the byte size of your records.
2) You can change the record type and not have to skip around in the code and change hard coded values later.

[Dimster]

I see - so how then would a REDIM work?

Type EventN
  Redim LastEvent as Double
  Redim etc
  EventDec as string * 25
End Type

or Redim doesn't work with UDT's, would need to create and whole new Type?

[bplus]

I see - so how then would a REDIM work?

Type EventN
  Redim LastEvent as Double
  Redim etc
  EventDec as string * 25
End Type

or Redim doesn't work with UDT's, would need to create and whole new Type?

REDIM works great for UDT's when you want a Dynamic array that can change it's size, say when you want to add more records to an array of records.

REDIM myEvents(1 to 20) as EventN  'and now you have 20 records ready to be filled with data or loaded from a file.

Use DIM myArray(1 to 20) 'a Static array, when you know for sure you don't have to change size of array, I think it's a bit faster processing things when it's Static. Now days not a big difference until you get into huge arrays.

[Dimster]

Thanks very much Mark. With Steve's help and yours I have much better understanding of the How To's for Random access files.
It's very, very much appreciated.

[bplus]

Hi Dimster,

My quess is that you've not used UDT arrays yet? Brave New World!

To access the .LastEvent element in your 5th record:

Code: QB64: [Select]

1. REDIM myArrayName(1 to 20) as  eventN
2. ...
3. LE5th = myArrayName(5).LastEvent

or to set 10th record .LastEvent element

Code: QB64: [Select]

1. myArrayName(10).LastEvent = x!

or set all .LastEvent elements to 100

Code: QB64: [Select]

1. FOR i = LBOUND(myArrayName) to UBOUND(myArrayName)
2.    myArrayName(i).LastEvent = 100
3. NEXT

   

[Dimster]

You are partially correct. I know of them and have seen coding using them but I tended to avoid them like the plague. I have found for accuracy of data which undergoes multiple calculations and updates, the simpler I kept the arrays the better accuracy results. But now I find I'm at the point where I need to explore Random Access of the data and UDT's, as confusing as they are for me, are in fact that Brave New World (you nailed it).

I will go back over this forum's history and review everything about UDT's in QB64.

Thanks for info on working within the elements of a UDT array. It will be very handy.

[SMcNeill]

A quick and simple Random Access File demo: 

Code: QB64: [Select]

1. _TITLE "Random Access File Demo"
2.  
3. TYPE Person_Type 'a type to hold your records
4.     FirstName AS STRING * 20
5.     LastName AS STRING * 20
6.     Phone AS _UNSIGNED _INTEGER64
7. END TYPE
8.  
9. DIM PhoneRecord AS Person_Type 'and a variable to reference that type
10.  
11. OPEN "Phone Database" FOR RANDOM AS #1 LEN = LEN(PhoneRecord) 'open file at the length of one of those records.
12.  
13. DO
14.     CLS
15.     RecordLimit = LOF(1) \ LEN(PhoneRecord) 'the number of records in your file.
16.     PRINT "Record"; TAB(5); "Last Name"; TAB(30); "First Name"; TAB(55); "Phone Number"
17.     FOR i = 1 TO RecordLimit
18.         GET #1, i, PhoneRecord 'get the whole record
19.         PRINT i; TAB(5); PhoneRecord.LastName; TAB(30); PhoneRecord.FirstName; TAB(55); PhoneRecord.Phone
20.     NEXT
21.     PRINT
22.     PRINT "What to do: 1)Add a record"
23.     PRINT "            2)Delete a record"
24.     PRINT "            3)Edit a record"
25.     PRINT "            4)QUIT"
26.     DO
27.         a = VAL(INPUT$(1))
28.     LOOP UNTIL a > 0 AND a < 5
29.     PRINT a
30.     PRINT
31.  
32.     SELECT CASE a
33.         CASE 1 'add a record
34.             RecordLimit = RecordLimit + 1
35.             PRINT "Adding Record #"; RecordLimit
36.             INPUT "Enter Last Name:"; PhoneRecord.LastName
37.             INPUT "Enter First Name:"; PhoneRecord.FirstName
38.             INPUT "Enter Phone Number:"; PhoneRecord.Phone
39.             PUT #1, RecordLimit, PhoneRecord
40.         CASE 2 'delete a record
41.             INPUT "Which record to delete?"; DeleteRecord
42.             FOR i = DeleteRecord TO RecordLimit - 1 'shift all records from deletion point on back one spot
43.                 GET #1, i + 1, PhoneRecord
44.                 PUT #1, i, PhoneRecord
45.             NEXT
46.             CLOSE #1
47.             a$ = SPACE$(LEN(PhoneRecord) * (RecordLimit - 1)) 'soace for all but the last record
48.             OPEN "Phone Database" FOR BINARY AS #1
49.             GET #1, 1, a$ 'get the whole database (minus the deleted record)
50.             CLOSE #1
51.             OPEN "Phone Database" FOR OUTPUT AS #1: CLOSE 'delete the old database
52.             OPEN "Phone Database" FOR BINARY AS #1 'save the new database to the drive.
53.             PUT #1, 1, a$
54.             CLOSE #1
55.             OPEN "Phone Database" FOR RANDOM AS #1 LEN = LEN(PhoneRecord) 'open file at the length of one of those records.
56.         CASE 3 'edit a record
57.             INPUT "Which Record to edit?"; editrecord
58.             IF editrecord > 0 AND editrecord <= RecordLimit THEN
59.                 PRINT "Editing Record #"; editrecord
60.                 INPUT "Enter Last Name:"; PhoneRecord.LastName
61.                 INPUT "Enter First Name:"; PhoneRecord.FirstName
62.                 INPUT "Enter Phone Number:"; PhoneRecord.Phone
63.                 PUT #1, editrecord, PhoneRecord
64.             ELSE
65.                 PRINT "Invalid record number."
66.             END IF
67.         CASE 4 'system
68.     END SELECT
69.  
70. LOOP
71.  

[Dimster]

Steve - if we assume the person has more than 1 phone number. It would be the Type Person_Type that would handle this, correct? And if so is the change:

Phone (1 to 3) AS _UNSIGNED _INTEGER64

or

Phone1 AS _UNSIGNED _INTEGER64
Phone2 AS _UNSIGNED _INTEGER64
Phone3 AS _UNSIGNED _INTEGER64

[bplus]

Steve - if we assume the person has more than 1 phone number. It would be the Type Person_Type that would handle this, correct? And if so is the change:

Phone (1 to 3) AS _UNSIGNED _INTEGER64

or

Phone1 AS _UNSIGNED _INTEGER64
Phone2 AS _UNSIGNED _INTEGER64
Phone3 AS _UNSIGNED _INTEGER64

Hi Dimster,

I am not Steve but can answer your question now:

The first will not work as arrays can't be used in types. The 2nd will work.

There is a 3rd option, caution might be confusing but WTH might be clarifying?

You could setup another TYPE

Code: QB64: [Select]

1. TYPE phonesType
2.   Home AS _UNSIGNED _INTEGER64
3.   AltHome AS _UNSIGNED _INTEGER64
4.   Cell AS _UNSIGNED _INTEGER64
5.   AltCell AS _UNSIGNED _INTEGER64
6.   Emergency AS _UNSIGNED _INTEGER64
7.   AltEmergency AS _UNSIGNED _INTEGER64
8. END Type
9.  
10. 'and install this type in your Person_Type
11. TYPE Person_Type 'a type to hold your records
12.     FirstName AS STRING * 20
13.     LastName AS STRING * 20
14.     Phones AS phonesType
15. END TYPE

[Dimster]

OH, a TYPE within a TYPE - clever Mark, clever