◄Contents►  ◄Index►  ◄Back►
─────────────────────────────────────────────────────────────────────────────
                           ◄Knowledge Base Contents►  ◄Knowledge Base Index►
 
 ISAMIO /E Cannot Extract Aggregate Types from ISAM Database - Q66139
 
 The ISAMIO.EXE utility can extract the data from an ISAM table into an
 ASCII text file by using the /E switch. However, the ISAMIO.EXE
 utility cannot completely extract aggregate items in a table.
 Aggregate items are items in the TYPE record that are not simple
 types. Simple types are INTEGER, LONG, CURRENCY, DOUBLE, and STRING.
 Aggregate types are items such as nested TYPEs or an array in the
 TYPE. This is not a problem with the ISAMIO.EXE utility, but is a
 result of the way in which the ISAM engine stores aggregate types in
 the table and in the data dictionary. This article also describes two
 ways to view a database's data dictionary.
 
 More Information:
 
 This behavior of ISAMIO results from the way in which the ISAM engine
 stores an aggregate item in the database. For an aggregate type, it
 calculates the size of the whole aggregate item, and then inserts an
 entry in the data dictionary describing it as a binary object of that
 many bytes. For example, consider the following TYPE:
 
 TYPE nested ' This gets nested in the TYPE below.
     junk1 AS DOUBLE
     junk2 AS STRING * 7
 END TYPE
 
 TYPE record
    key1 AS STRING * 20
    key2 AS nested
    key3 AS INTEGER
    key4(5) AS INTEGER
    key5 AS LONG
    key6(4)  AS DOUBLE
    key7 AS DOUBLE
    key8 AS CURRENCY
 END TYPE
 
 If you create a table in a database with this TYPE, the ISAM engine
 will put entries in the data dictionary stating that key2, key4, and
 key6 are "variablestring" objects of a certain size. Thus, when
 ISAMIO later tries to extract these items, it cannot determine what
 their original data types were; it only knows how big they were.
 Therefore, ISAMIO extracts these aggregate objects as a string
 (usually full of graphics characters).
 
 You can show what is in a database's data dictionary in two ways, as
 described below:
 
 1. ISAMIO can access the data dictionary. When you run ISAMIO to
    extract a database as a text file, you can specify a <specfile>
    parameter. For example, to extract the table called "test" in a
    database called "test.mdb" made with the above TYPEs, you can use
    the following line:
 
       ISAMIO /E data.txt test.mdb test specfile.txt /C
 
    After this, DATA.TXT contains the output data. While it's
    extracting, ISAMIO takes the information in the data dictionary and
    builds "SPECFILE.TXT". Note that SPECFILE.TXT does not have to exist
    before you run ISAMIO. After running ISAMIO, this specfile will
    describe the data items in the ASCII text file just created. For
    instance, for the TYPE example above, the specfile created is as
    follows:
 
    variabletext,20,key1
    variablestring,15,key2
    integer,,key3
    variablestring,12,key4
    long,,key5
    variablestring,40,key6
    double,,key7
    currency,,key8
 
    This clearly shows key2, key4, and key4 are seen as
    'variablestring' objects by ISAMIO.EXE.
 
 2. The second way to examine the data dictionary is to run the
    ISAMPACK.EXE utility. If you redirect the output from ISAMPACK into
    a file, you will have a record of the database's contents. For
    instance, use the following:
 
       ISAMPACK test.mdb > report.dat
 
    This creates a file called REPORT.DAT, which (among other things)
    has a description of the structure of each table in the database.
    The structure of the tables is determined by what is in the data
    dictionary. For the table discussed above, a part of REPORT.DAT
    will contain the following:
 
       Column Name               Column Type         Maximum Size
       -----------               -----------         ------------
 
         key1                      VarText             20
         key2                      VarString           15
         key3                      Integer             2
         key4                      VarString           12
         key5                      Long                4
         key6                      VarString           40
         key7                      Double              8
         key8                      Currency            8
 
    Again, this shows that the data dictionary describes aggregate
    objects as being "VarString".
 
    See the following example to build the code example described
    above.
 
 Code Example
 ------------
 
 To build the code example described above, you can use the following
 program. To run this program in VBDOS.EXE, you must first load the
 PROISAMD.EXE TSR. To compile and run the program, use the following:
 
    BC test.bas ;
    LINK test ;
 
 DEFINT A-Z
 TYPE nested
     junk1 AS DOUBLE
     junk2 AS STRING * 7
 END TYPE
 
 TYPE record
    key1 AS STRING * 20
    key2 AS nested
    key3 AS INTEGER
    key4(5) AS INTEGER
    key5 AS LONG
    key6(4)  AS DOUBLE
    key7 AS DOUBLE
    key8 AS CURRENCY
 END TYPE
 
 DIM Record1 AS record, Record2 AS record, Record3 AS record
 ' Code part.
 
 OPEN "test.mdb" FOR ISAM record "test" AS #1
 ' SETINDEX #1, "testindex"
 
 FOR i = 1 TO 20
    PRINT "***********************************"
    PRINT "*       Get a new record          *"
    PRINT "***********************************"
 
    INPUT "Input a STRING * 20: ", Record1.key1
    INPUT "Input a DOUBLE: ", Record1.key2.junk1
    INPUT "Input a STRING * 7: ", Record1.key2.junk2
    INPUT "Input an INTEGER: ", Record1.key3
    FOR j = 1 TO 5
      INPUT "Input an INTEGER: ", Record1.key4(j)
    NEXT
    INPUT "Input a LONG: ", Record1.key5
    FOR j = 1 TO 4
      INPUT "Input an INTEGER: ", Record1.key6(j)
    NEXT
    INPUT "Input a DOUBLE: ", Record1.key7
    INPUT "Input a CURRENCY: ", curr1
    INSERT #1, Record1
 NEXT
 CLOSE
 END