Big-Data-HPC-Platform/docs/ECLLanguageReference/ECLR_mods/Basics-FieldandAttribute.xml

<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE sect1 PUBLIC "-//OASIS//DTD DocBook XML V4.5//EN"
"http://www.oasis-open.org/docbook/xml/4.5/docbookx.dtd">
<sect1 id="Field_and_Attribute_Qualification">
  <title>Field and Definition Qualification</title>

  <sect2 id="Imported_Attributes">
    <title>Imported Definitions</title>

    <para>EXPORTed<indexterm>
        <primary>EXPORTed</primary>
      </indexterm> definitions<indexterm>
        <primary>EXPORTed Definitions</primary>
      </indexterm> defined within another module and IMPORTed<indexterm>
        <primary>IMPORTed</primary>
      </indexterm> (see the <link linkend="EXPORT">EXPORT</link> and <link linkend="IMPORT">IMPORT</link>
     keywords) are available for use in
    the definition that contains the IMPORT. Imported Definitions must be
    fully qualified by their Module name and Definition name, using dot
    syntax<indexterm>
        <primary>dot syntax</primary>
      </indexterm> (module.definition).</para>

    <programlisting>IMPORT abc;                //make all exported definitions in the abc module available
EXPORT Definition1 := 5;    //make Definition1 available to other modules
Definition2 := abc.Definition2 + Definition1;  // object qualification needed for Definitions from abc module</programlisting>
  </sect2>

  <sect2 id="Fields_in_Datasets">
    <title>Fields in Datasets</title>

    <para>Each Dataset counts as a qualified scope and the fields within them
    are fully qualified by their Dataset (or record set) name and Field name,
    using dot syntax (dataset.field). Similarly, the result set of the
    TABLE<indexterm>
        <primary>TABLE</primary>
      </indexterm> built-in function (see the <emphasis
    role="bold">TABLE</emphasis> keyword) also acts as a qualified scope. The
    name of the record set to which a field belongs is the object name:</para>

    <programlisting>Young := YearOf(Person.per_dbrth) &lt; 1950;
MySet := Person(Young);</programlisting>

    <para>When naming a Dataset as part of a definition, the fields of that
    Definition (or record set) come into scope. If Parameterized Definitions
    (functions) are nested, only the innermost scope is available. That is,
    all the fields of a Dataset (or derived record set) are in scope in the
    filter expression. This is also true for expressions parameters of any
    built-in function that names a Dataset or derived record set as a
    parameter.</para>

    <programlisting>MySet1 := Person(YearOf(dbrth) &lt; 1950);
// MySet1 is the set of Person records who were born before 1950</programlisting>

    <para></para>

    <programlisting>MySet2 := Person(EXISTS(OpenTrades(AgeOf(trd_dla) &lt; AgeOf(Person.per_dbrth))));</programlisting>

    <para></para>

    <programlisting>// OpenTrades is a pre-defined record set.
//All Trades fields are in scope in the OpenTrades record set filter 
//expression, but Person is required here to bring Person.per_dbrth
// into scope
//This example compares each trades' Date of Last Activity to the
// related person’s Date Of Birth</programlisting>

    <para>Any field in a Record Set <emphasis role="underline">can</emphasis>
    be qualified with either the Dataset name the Record Set is based on, or
    any other Record Set name based on the same base dataset. For
    example:</para>

    <programlisting>memtrade.trd_drpt
nondup_trades.trd_drpt
trades.trd_drpt</programlisting>

    <para>all refer to the same field in the memtrade dataset.</para>

    <para>For consistency, you should typically use the base dataset name for
    qualification. You can also use the current Record Set's name in any
    context where the base dataset name would be confusing.</para>
  </sect2>

  <sect2 id="Scope_Resolution_Operator">
    <title>Scope Resolution Operator</title>

    <para>Identifiers are looked up in the following order:</para>

    <para>1. The currently active dataset, if any</para>

    <para>2. The current definition being defined, and any parameters it is
    based on</para>

    <para>3. Any definitions or parameters of any MODULE or FUNCTION structure
    that contains the current definition</para>

    <para>This might mean that the definition or parameter you want to access
    isn't picked because it is hidden as in a parameter or private definition
    name clashing with the name of a dataset field.</para>

    <para>It would be better to rename the parameter or private definition so
    the name clash cannot occur, but sometimes this is not possible.</para>

    <para>You may direct access to a different match by qualifying the field
    name with the scope resolution operator (the carat (^) character), using
    it once for each step in the order listed above that you need to
    skip.</para>

    <para>This example shows the qualification order necessary to reach a
    specific definition/parameter:</para>

    <para><programlisting>ds := DATASET([1], { INTEGER SomeValue });

INTEGER SomeValue := 10; //local definition

myModule(INTEGER SomeValue) := MODULE

  EXPORT anotherFunction(INTEGER SomeValue) := FUNCTION
    tbl := TABLE(ds,{SUM(GROUP, someValue), // 1 - DATASET field
                     SUM(GROUP, ^.someValue), // 84 - FUNCTION parameter
                     SUM(GROUP, ^^.someValue), // 42 - MODULE parameter
                     SUM(GROUP, ^^^.someValue), // 10 - local definition
                 0});
     RETURN<indexterm>
          <primary>RETURN</primary>
        </indexterm> tbl;
   END;

   EXPORT result := anotherFunction(84);
   END;

OUTPUT(myModule(42).result);</programlisting></para>

    <para>In this example there are four instances of the name
    "SomeValue":</para>

    <para>a field in a DATASET.</para>

    <para>a local definition</para>

    <para>a parameter to a MODULE structure</para>

    <para>a parameter to a FUNCTION structure</para>

    <para>The code in the TABLE function<indexterm>
        <primary>TABLE function</primary>
      </indexterm> shows how to reference each separate instance.</para>

    <para>While this syntax allows exceptions where you need it, creating
    another definition with a different name is the preferred solution.</para>
  </sect2>
</sect1>