BlueIron

BlueIron is a library that takes a set of rows as input, applies some transformations and output a new set of rows. For instance, one can sort the input rows and create some breaks with title and total rows.

Consider the following input rows:

With the following BlueIron descriptor:

<blueiron>
  <!-- input data structure -->
  <input id="employees">
    <column id="ID" type="String" />
    <column id="FIRSTNAME" type="String" />
    <column id="LASTNAME" type="String" />
    <column id="SEX" type="String" />
    <column id="BIRTHDATE" type="Date" />
    <column id="DEPARTMENT" type="String" />
    <column id="SALARY" type="BigDecimal" />
  </input>

  <!-- global step holder -->
  <step id="final" source="employees">
    <!-- a break per departement with a total line. Inner data are thrown away. -->
    <break data="false">
      <condition type="column" id="DEPARTMENT" />
      <total>
        <copy id="DEPARTMENT" />
        <copy id="SALARY" source="SUM_SALARY" />
      </total>
      <computations>
        <expr id="SUM_SALARY">previousValue + SALARY</expr>
      </computations>
    </break>
    <!-- keep only the 2 last columns -->
    <project>
      <copy id="DEPARTMENT" />
      <copy id="SALARY" />
    </project>
  </step>
</blueiron>

Will produce the following output:

In order to provide maximum flexibility in data transformation, BlueIron uses the concept of step. Those steps will then use one or more processors.

A step denotes a unit of work: it reads the input rows, makes its transormation(s) and then output a (potentially) new set of rows. Then many steps will be piped together to produce the final output result.

Generally, a step is needed especially in this cases:

• The structure of the output is changed
• New rows are created
• Rows are removed

BlueIron comes with a set of basic steps which