Processing Filters

These filters are available for use with Kafka Connect File Pulse:

AppendFilter

The following provides usage information for : io.streamthoughts.kafka.connect.filepulse.filter.AppendFilter

The AppendFilter is probably one of the most important processing filters to know. It allows you to manipulate a source record by easily adding or replacing a field with a constant value or a value extracted from another existing field using ScEL.

Configuration

Examples

The following examples shows how to use the AppendFilter to concat two values from the array field named values using a substitution expression. The concat value is then added to the a field named result.

Configuration

filters.SubstituteFilter.field="result"
filters.SubstituteFilter.value="{{ extract_array(values,0) }}-{{ extract_array(values,1) }}"

Input

{
    "record" :{
      "value": ["Hello", "World"]
    }
} 

Output

{
    "record" :{
      "value": ["Hello", "World"],
      "result" : "Hello-World"
    }
}

In the previous example, we used the simple property expression result to indicate the target field to which our value is added. We have actually omitted the expression scope $value. By default, if no scope is defined in an expression, the scope $value is implicitly applied. Hence, we could have used the fully expression $value.result which is similar to the simplified expression result.

But, you can perfectly used another expression scope. For example, you can leverage the AppendFilter to dynamically resolve the record-key or the output topic based on the record data.

The following configuration show how to use the $topic scope :

filters.SubstituteFilter.field="$topic"
filters.SubstituteFilter.value="my-topic-{{ lowercase(extract_array(values,0) }}"

Input

{
  "record" : {
    "value": ["Hello", "World"]
  }
} 

Output

{
  "context": { "topic" : "my-topic-hello" },
  "record" :{
    "value": ["Hello", "World"]
  }
} 

Finally the AppendFilter can also accept an substitution expression for the property field. This allows to dynamically determine the name of the field to be added.

The following examples show how to use a property expression to get the named of the field from a

filters.SubstituteFilter.field="{{ target }}"
filters.SubstituteFilter.value="{{ extract_array(values,0) }}-{{ extract_array(values,1) }}"

Input

{
  "record" : {
    "target": "result",
    "value": ["Hello", "World"]
  }
}

Output

{
  "record" : {
    "target": "result",
    "value": ["Hello", "World"],
    "result" : "Hello-World"
  }
}

ConvertFilter

The following provides usage information for : io.streamthoughts.kafka.connect.filepulse.filter.ConvertFilter

The ConvertFilter can be used to convert a field’s value into a specific type.

Configuration

Supported types are :

• SHORT
• INTEGER
• LONG
• FLOAT
• DOUBLE
• BOOLEAN
• STRING
• ARRAY
• BYTES

Examples

The following example shows how to convert a a field’s value containing the string yes into a boolean.

Configuration

filters.BooleanConverter.field="target"
filters.BooleanConverter.type="BOOLEAN"

Input

{
  "record" : {
    "target": "yes"
  }
}

Output

{
  "record" : {
    "target": true
  }
}

DateFilter

The following provides usage information for : io.streamthoughts.kafka.connect.filepulse.filter.DateFilter

The DateFilter converts a field’s value containing a date to a unix epoch time.

Configuration

Examples

filters.MyDateFilter.field="date"
filters.MyDateFilter.target="timestamp"
filters.MyDateFilter.format="yyyy-MM-dd'T'HH:mm:ss"

Input

{
  "record" : {
    "date": "2001-07-04T12:08:56"
  }
}

Output

{
  "record" : {
    "date": "2001-07-04T12:08:56",
    "timestamp": 994248536000
  }
}

DelimitedRowFilter

The following provides usage information for : io.streamthoughts.kafka.connect.filepulse.filter.DelimitedRowFilter.

The DelimitedRowFilter can be used to parse and stream delimited row files (i.e CSV) into Kafka. Each row is parsed and published into a configured topic as a single Kafka data.

Configuration

Examples

The following example shows the use of the DelimitedRowFilter to split the message field using | as a separator character. The name of each column is extracted from the fields headers.

filters.ParseDelimitedRow.extractColumnNam="headers"
filters.ParseDelimitedRow.separator="\\|"
filters.ParseDelimitedRow.trimColumn="true"
filters.ParseDelimitedRow.type="io.streamthoughts.kafka.connect.filepulse.filter.DelimitedRowFilter"

DropFilter

The following provides usage information for : io.streamthoughts.kafka.connect.filepulse.filter.DropFilter.

The DropFilter can be used to prevent some messages (i.e records) to be written into Kafka.

Configuration

For more information about if property, see : Conditional execution.

Examples

The following example shows the usage of DropFilter to only keep records with a field level containing to ERROR.

filters=Drop
filters.Drop.type=io.streamthoughts.kafka.connect.filepulse.filter.DropFilter
filters.Drop.if={{ equals(level, ERROR) }}
filters.Drop.invert=true

FailFilter

The following provides usage information for : io.streamthoughts.kafka.connect.filepulse.filter.FailFilter.

The fail filter can be used to throw an exception with a provided error message. For example, this can be useful to stop processing a file when a non-conform record is read.

Configuration

Examples

The following example shows the usage of FailFilter to stop processing a file when a field is equals to null.

filters=Fail
filters.Fail.type=io.streamthoughts.kafka.connect.filepulse.filter.FailFilter
filters.Fail.if={{ is_null(user_id) }}
filters.Fail.message=Invalid row, user_id is missing : {{ $value }}

GrokFilter

The following provides usage information for : io.streamthoughts.kafka.connect.filepulse.filter.GrokFilter.

The GrokFilter allows you to parse unstructured data like applications logs to extract structured and meaningful data fields.

Regular Expressions Grok are built on top of on regular expressions, so you can use any regular expressions as well to define your own patterns.

Internally, FilePulse uses the regular expression library is Joni , a Java port of Oniguruma regexp library.

Configuration

Examples

The following example shows the usage of GrokFilter to parse and extract fields from application log message.

filters.ParseLog4jLog.match="%{TIMESTAMP_ISO8601:logdate} %{LOGLEVEL:loglevel} %{GREEDYDATA:message}"
filters.ParseLog4jLog.overwrite="message"
filters.ParseLog4jLog.source="message"
filters.ParseLog4jLog.type="io.streamthoughts.kafka.connect.filepulse.filter.GrokFilter"
filters.ParseLog4jLog.ignoreFailure="true"

GroupRowFilter

The following provides usage information for : io.streamthoughts.kafka.connect.filepulse.filter.GroupRowFilter.

Configuration

Examples

JoinFilter

The following provides usage information for : io.streamthoughts.kafka.connect.filepulse.filter.JoinFilter.

Configuration

Examples

JSONFilter

The following provides usage information for : io.streamthoughts.kafka.connect.filepulse.filter.JSONFilter.

The JSONFilter parses an input json field.

Configuration

Examples

filters=MyJsonFilter
filters.MyJsonFilter.type=io.streamthoughts.kafka.connect.filepulse.filter.JSONFilter
filters.MyJsonFilter.source=message
filters.MyJsonFilter.target=payload

MultiRowFilter

The following provides usage information for : io.streamthoughts.kafka.connect.filepulse.filter.MultiRowFilter.

The MultiRowFilter joins multiple lines into a single Struct using a regex pattern.

Configuration

Examples

The following example shows the usage of the MultiRowFilter to join Java exception message with its stacktrace.

filters=StackTraceMultiRowFilter
filters.StackTraceMultiRowFilter.type=io.streamthoughts.kafka.connect.filepulse.filter.MultiRowFilter
filters.StackTraceMultiRowFilter.negate=false
filters.StackTraceMultiRowFilter.pattern=^[\t]

RenameFilter

The following provides usage information for : io.streamthoughts.kafka.connect.filepulse.filter.RenameFilter.

The RenameFilter is used to rename a specified field.

Configuration

Examples

filters.MyJsonFilter.field=input
filters.MyJsonFilter.target=renamed

Input

{
  "record" : {
    "input": "foo"
  }
}

Output

{
  "record" : {
    "renamed": "foo"
  }
}

SplitFilter

The following provides usage information for : io.streamthoughts.kafka.connect.filepulse.filter.SplitFilter.

The SplitFilter splits a field’s value of type string into an array by using a specific separator.

Configuration

Example

Configuration

filters.MySplitterFilter.split=input
filters.MySplitterFilter.separator=,

Input

{
  "record" : {
    "input": "val0,val1,val2"
  }
}

Output

{
  "record" : {
    "input": "val0,val1,val2",
    "output": [ "val0", "val1", "val2"]
  }
}