JDBC Importer

JDBC Importer Tutorial 10 : One Table, CSV Delimited, New Row Translator

Please make sure you have the appropriate libraries in your classpath (including the JDBC driver used to connect to your database) before starting the tutorials.

In this tutorial, you'll learn the basics of creating a new Row Translator and running the import with it. The table that will contain the rows imported is called employee and it has the following columns:

Make sure that these table(s) are created in the database that you'll be importing data. You can find the oracle creation script in the samples directory under the filename : 'tutorial10/createtable_ora.sql'.

Now that the database is setup, you can examine the architecture to see how the Row Translator is used during the import.

Architecture Background

The Row Translator is used during the import for converting the row's values read from the file by adding/removing columns, changing column values or skipping the entire row.

The JDBCImporter requires that each Row Translator be a Java Bean like object. All <property> tags defined inside the '<translator>' will be passed to the appropriate set method. The RowTranslator must also implement one method : getValues().

Custom Row Translator

The Row Translator that you will be creating will skip over the first n rows. The number of rows to skip will be specified in a property called 'numRows'.

The first thing to do is create the class SkipRowsRowTranslator that implements the RowTranslator interface.

import net.sourceforge.jdbcimporter.EntityDef;
import net.sourceforge.jdbcimporter.RowTranslator;
import net.sourceforge.jdbcimporter.ColumnValue;

public class SkipRowsRowTranslator implements RowTranslator 
{
  public ColumnValue[] getValues(EntityDef entity, ColumnValue[] rowValues) 
  {
    /* @todo implement this method */
    return rowValues;
  }
}
				

The Row Translator needs two properties: the number for rows to skip and the current row:

...

public class SkipRowsRowTranslator implements RowTranslator 
{
  protected int numRows = 0;
  protected int curRow  = 0;
  
  public void setNumRows( int numRows )
  {
    this.numRows = numRows;
  }
  
  ...
  
}
				

The number of rows to skip property will be set before the getValues method is called and the current row is intialized to zero.

The SkipRowsRowTranslator first checks if the current row is less than the number of rows to skip. If this is true then null is returned which will indicate that the current row should be skipped. Otherwise, the row's values will be returned. In each case, the current row property is incremented.

...

public class SkipRowsRowTranslator implements RowTranslator 
{
  ...
  
  public ColumnValue[] getValues(EntityDef entity, ColumnValue[] rowValues) 
  {
    if ( curRow < numRows )
    { 
      curRow++;
      return null;
    }
    else 
    {
      curRow++;
      return rowValues;
    }
  }
}
				

This ends the tutorial for creating the custom Row Translator. The full source code of the SkipRowsRowTranslator is found under the package 'samples.rowtranslator'. What follows now is the instructions on how to use the custom Row Translator during the import. If you have read through the first tutorial then you may wish to skip to the row translator section. The other sections are the same as the first tutorial.

Import Config XML

Now that the database is setup, you can examine the import XML config file that will be used (in the samples directory under the filename : 'tutorial10/import.xml'). The file begins with the standard XML document declaration followed by the '<import>' tag. This tag indicates that there is an import to be processed. There are seven attributes specified on the '<import>' tag: the 'log' attribute, the 'bad' attribute, the 'commitCount', the 'batchCount' attribute, the 'preSQLFile' attribute, the 'postSQLFile' attribute and the 'trimValues' attribute. The 'log' attribute specifies a filename into which JDBCImporter writes all audit, error, and warnings that occur during the import. The 'bad' attribute specifies a filename into which JDBCImporter writes data that was not properly imported into the database. The 'commitCount' attribute specifies how many rows to import before calling commit on the JDBC Connection. The 'batchCount' attribute specifies how many rows to import before calling executeBatch on the import engine (when the JDBC driver supports batch mode). By default, the 'commitCount' and 'batchCount' attributes are set 1, auto commit is turned on and batch mode is not used. The 'preSQLFile' and the 'postSQLFile' attributes specify filenames that contain sql statements to be executed before and after the import , respectively. The 'trimValues' attribute specifies whether strings values read from the Delimiter Parser are trimmed (ie. remove leading and trailing whitespace). By default, it is set to false.

There are two parts inside the '<import>' tag that define how and where the data is imported: the connection definition and the entity definitions.

Connection Definition

The connection definition begins with '<connection>' tag and contains the information needed to connect to the database. In this tutorial, you will be using the JDBC DriverManager to initialize a connection to the database. To indicate this, the 'type' attribute's value, inside the '<connection>' tag, is 'jdbc'. The specific connection information is found inside the '<connection>' tag as '<property>' tags. A '<property>' tag has two attributes: 'name' specifies the name of the property and 'value' specifies the string value of the property. For the JDBC DriverManager, you will need to specify the following information: the driver class name (with the property name 'driver'), the connection url (with the property name 'url'), the username (with the property name 'username'), the password (with the property name 'password'). The following is an example of the connection definition :

 <connection type="jdbc"> 
    <property name="driver" value="oracle.jdbc.driver.OracleDriver"/> 
    <property name="url" value="jdbc:oracle:thin:@localhost:1521:orcl"/> 
    <property name="username" value="scott"/> 
    <property name="password" value="tiger"/> 
 </connection> 

Entity Definition

Since you will be importing data into one table, there will be only one entity definition.In general, you will need an entity definition for each table that you will be importing data. Remember to specify the entity definitions in the order that the import should occur. For example, if table 'ingredient' depends on table 'recipe' (ie. has a foreign key), the entity definition of table 'recipe' should be placed before the entity definition of table 'ingredient'. Every entity definition begins with '<entity>' tag.

The 'table' attribute must contain the name of the table. Optionally, you can further specify the table by providing values for the 'schema' and the 'catalog' attributes.

To specify a custom import engine to process the entity, you may add the 'engine' attribute, whose value is the classname of the import engine.In this tutorial, you will be using the default import engine.

The 'source' attribute must contain the data file location.From looking at the sample data (found under 'samples/tutorial5/employee.csv'), you will see that there are 8 columns that are separated by the ',' character.

There are three parts inside the '<entity>' tag : the delimiter parser definition, row translator definition, and the list of columns found in the data file.

Delimiter Parser

The delimiter parser definition begins with the '<delimiter>' tag and contains the information needed to parse the input file into a set of rows that will be imported into the table.In this tutorial, you will be using the CSV Delimiter Parser. To indicate this, the 'type' attribute's value, inside the '<delimiter>' tag, is 'csv'. The specific Delimiter Parser information is found inside the '<delimiter>' tag. For the CSV Delimiter Parser, you will need to specify the following information (as '<property>' tags): the string that delimits a column (in the property named 'columnDelimiter'), the string that encloses a column (optional, in the property named 'enclosedDelimiter'), whether the string that encloses a column is optional (in the property named 'enclosedOptional', it must have a value of 'true' or 'false'). Since, the data file has only a column delimiter (',' is the string separating the columns), the Delimiter Parser definition will look like this :

  <delimiter type="csv"> 
    <property name="columnDelimiter" value=","/> 
  </delimiter> 

Row Translator

The row translator definition is optional and begins with the '<translator>' tag. It contains the information needed to translate each row's values and may add, remove column values or skip the whole row. Since you will be using the custom Row Translator, you will have to specify the '<translator>' tag inside the '<entity>' tag (right after the '<delimiter>' tag). You must choose an identifier for the custom Row Translator (ex. 'tutorial_skipRows') and set the 'type' attribute's value, inside the '<translator>' tag, to that identifier. The specific Rows Translator information is found inside the '<translator>' tag. For the Skip Rows Row Translator, the number of rows to skip ('numRows') property should be set. To set the 'numRows' property to '5', the Row Translator definition would look like this:

 <translator type="tutorial_skipRows"> 
    <property name="numRows" value="5"/> 
 </translator> 

List of Columns

The final portion of the entity definition is the list of columns that are to be imported from the input file into the database. The list of columns should be the same order as they appear in the input file. Each column is defined inside the '<column>' tag. The name of the column must appear in the 'name' attribute of the '<column>' tag. Optionally, the java.sql.Type may be specified in the 'SQLType' attribute of the '<column>'.You will be letting the JDBC Importer figure out most of the column types (except for dates) in the database, so the 'SQLType' attribute is omitted except for the 'startdate' column. Here is an example of how the list of columns are defined in the import definition:

  <column name="id"></column>
  <column name="firstname"></column>
  <column name="lastname"></column>
  <column name="jobdescription"></column>
  <column name="manager"></column>
  <column name="startdate" SQLType="DATE"></column>

Running the Import

By now, the import definition should look like this (with your appropriate connection information):

<import log="import.log bad="import.bad"> 
  <connection type="jdbc"> 
     <property name="driver" value="oracle.jdbc.driver.OracleDriver"/> 
     <property name="url" value="jdbc:oracle:thin:@localhost:1521:orcl"/> 
     <property name="username" value="scott"/> 
     <property name="password" value="tiger"/> 
  </connection> 
  <entity table="employee" source="employee.csv">
    <delimiter type="csv"> 
      <property name="columnDelimiter" value=","/> 
    </delimiter> 
   <translator type="tutorial_skipRows"> 
      <property name="numRows" value="5"/> 
   </translator> 
    <column name="id"></column>
    <column name="firstname"></column>
    <column name="lastname"></column>
    <column name="jobdescription"></column>
    <column name="managerid"></column>
    <column name="startdate" SQLType="DATE"></column>
    <column name="salary"></column>
    <column name="department"></column>
  </entity> 
</import> 

Since you are using the custom Row Translator (Skip Rows Row Translator) you will have to create a property file with one entry that maps the identifier to the full name of the class that implements the RowTranslator interface. The entry's key should start with 'rowtranslator.' (this indicates that the custom component is a Row Translator). It should look like this:

rowtranslator.tutorial_skipRows=samples.rowtranslator.SkipRowsRowTranslator
				

You will also have to include an extra jar file in the classpath before you can use the custom Row Translator (the jar file 'jdbcimporter-samples.jar' under the directory 'lib' contains the custom Row Translator).

You can run the import by issuing the following command (assuming that the import definition and property file are in the current directory and are called 'import.xml' and 'custom.properties', respectively):

java net.sourceforge.jdbcimporter.Importer import.xml custom.properties

If all goes well then the two log files should be created. In the normal log file there should be an informational message indicating that all rows, except the first 5 rows, were imported. In the bad log file there should be a heading for the import table and the first 5 rows.