Populating Script Parameters at Runtime

When executing Hive scripts, you might want to populate their parameter values (e.g., username, password) at runtime. This way, you do not have to hard-code values in the Hive scripts and your implementation will be more dynamic, secure and easy to maintain. You can do this in two ways:

Let's take a look at each option.

Using the GUI to get values from the Registry

BAM provides a simple GUI for you to configure Hive scripts to get values from the Registry at runtime using placeholder keys with particular registry/repository prefixes. Only text/plain media type supports parameterization.

Follow the instructions below for a sample demonstration.

1. Log on to the BAM Management Console and select Registry-> Browse menu from the Main menu. That opens the Browse window. From there, select the Detail View tab.

2. Click the Add Resource link to add a resource.

Now, let's see how to configure the Hive script to get the value of this resource at runtime.

3. You can use the following notation to access a resource: ${conf:path} or ${gov:path} or ${local:path} from a Hive script. This is because the Registry consists of 3 repositories as follows:

• Local Data Repository  (local)
• Configuration Registry (conf)
• Governance Registry (gov)

The resource path you use in the script is prefixed with any of these 3 repositories depending on where you created the resource. It can be config, local or governance. For example, shown below is a resource by the name password in the local repository.

You can use : ${local:/repository/credential/password} value in that format for the password resource in the Hive script.

Extending a Java Class and implementing its getter and setter methods

Using the GUI as described above has some limitations since it only allows you to get values from the Registry at runtime. There can be instances where you have to dynamically get and set variable values that might not necessarily be stored in the Registry. For example, sometimes you have to perform a logical operation on data, store the result in a temporary variable and use that variable value in yet another operation. In such cases, you can write your own Java class by extending the AbstractHiveAnalyzer class and implementing its getter and setter methods.

The AbstractHiveAnalyzer class is shown below.

public abstract class AbstractHiveAnalyzer {
    public abstract void execute();
    public void setProperty(String key, String value) {
        HiveContext.getCurrentContext().setProperty(key, value);
    }
    public String getProperty(String key) {
        return HiveContext.getCurrentContext().getProperty(key);
    }
    public void removeProperty(String key, String value) {
        HiveContext.getCurrentContext().setProperty(key, null);
    }
}

You can use inherited property-methods to set parameters to the Hive configuration used during a particular Hive script execution. For example, if you set a property named lastRunTime using the setProperty method, the property will be available in subsequent Hive queries with notation {hiveconf:lasRunTime}. You can dynamically populate this lastRunTime property in each run of the script, by looking up values from the registry, databases etc. This is useful in maintaining state between two executions of a Hive script and using them subsequently.

In order to use the custom Java implementation, drop the JAR file containing the class to <BAM_HOME>/repository/components/lib and restart the server.

Use the following syntax in the implementation.

class 'Fully qualified class name of the implementation';

Generally, you can include this line of code in any place of the script according to your requirement. For example, if you want some pre-processing done, add it at the beginning of the script. If you want some post-processing to be done after normal Hive queries, add it at the end.

Also,  add the following code segment under maven dependencies.