ParallelReader

Not available in Community Designer

Short Description
Ports
Metadata
ParallelReader Attributes
Details
Examples
Best Practices
See also

Short Description

ParallelReader reads data from flat files using multiple threads.

Component Data source Input ports Output ports Each to all outputs Different to different outputs Transformation Transf. req. Java CTL Auto-propagated metadata
ParallelReaderflat file01-2
no
no
no
no
no
no
yes

Ports

Port typeNumberRequiredDescriptionMetadata
Output0
yes
for correct data recordsany
1
no
for incorrect data recordsspecific structure, see table bellow

Parsed data records are sent to the first output port.

The component has an optional output logging port for getting detailed information about incorrect records. To get all incorrect records together with the information about the incorrect value, its location, and the error message to error port, Data Policy has to be controlled and an edge has to be connected to the error port.

Metadata

ParallelReader has metadata template on the second output port.

Table 55.11. Error Metadata for Parallel Reader

Field NumberField ContentData TypeDescription
0record numberlongThe position of the erroneous record in the dataset (record numbering starts at 1)
1field numberintegerThe position of the erroneous field in the record (1 stands for the first field, i.e., that of index 0)
2original datastringThe erroneous record in raw form (including delimiters)
3error messagestringerror message - detailed information about this error
4reading thread offsetlongindicates the initial file offset of the parsing thread (optional field)

ParallelReader Attributes

AttributeReqDescriptionPossible values
Basic
File URL
yes
Data source(s) will be read. See Supported File URL Formats for Readers.  
Charset 

Encoding of records that are read in.

The default encoding depends on DEFAULT_CHARSET_DECODER in defaultProperties.

UTF-8 | <other encodings>
Data policy Determines what should be done when an error occurs. For more information, see Data Policy. Strict (default) | Controlled | Lenient
Trim strings specifies whether leading and trailing whitespace should be removed from strings before setting them to data fields, see Trimming Data. If true, the use of the robust parser is forced. false (default) | true
Quoted strings 

Fields containing a special character (comma, newline, or double quote) have to be enclosed in quotes. Only single/double quote is accepted as the quote character. If true, special characters are removed when read by the component (they are not treated as delimiters).

Example: To read input data "25"|"John", switch Quoted strings to true and set Quote character to ". This will produce two fields: 25|John.

By default, the value of this attribute is inherited from metadata on output port 0. See also Record Details.

false | true
Quote character

Specifies which kind of quotes will be permitted in Quoted strings. By default, the value of this attribute is inherited from metadata on output port 0. See also Record Details.

both | " | '
Advanced
Skip leading blanks 

Specifies whether to skip a leading whitespace (blanks e.g.) before setting input strings to data fields. If not explicitly set (i.e. having the default value), the value of the Trim strings attribute is used. See Trimming Data. If true, the use of the robust parser is enforced.

false (default) | true
Skip trailing blanks 

Specifies whether to skip a trailing whitespace (blanks e.g.) before setting input strings to data fields. If not explicitly set (i.e. having the default value), the value of the Trim strings attribute is used. See Trimming Data. If true, the use of the robust parser is enforced.

false (default) | true
Max error count 

The maximum number of tolerated error records in input file(s); applicable only if Controlled Data Policy is set

0 (default) - N
Treat multiple delimiters as one 

If a field is delimited by a multiplied delimiter char, it will be interpreted as a single delimiter when setting to true.

false (default) | true
Verbose 

By default, less comprehensive error notification is provided and the performance is slightly higher. However, if switched to true, more detailed information with less performance is provided.

false (default) | true
Level of parallelism 

Number of threads used to read input data files. The order of records is not preserved if it is 2 or higher. If the file is too small, this value will be switched to 1 automatically.

2 (default) | 1-n
Distributed file segment reading 

In case the component is running in a CloverETL Server Cluster environment and a shared file is read, each component's instance process the appropriate part of the file. The whole file is divided into segments by CloverETL Server and each cluster worker processes only one proper part of the file. By default, this option is turned off. This attribute is ignored for partitioned files.

false (default) | true
Parser 

By default, the most appropriate parser is applied. Besides, the parser for processing data may be set explicitly. If an improper one is set, an exception is thrown and the graph fails. See Data Parsers

auto (default) | <other>

Details

ParallelReader reads delimited flat files like CSV, tab delimited, etc., fixed-length, or mixed text files. The component can read a single file as well as a collection of files placed on a local disk or remotely, remote files are accessible via FTP and S3 protocol.

Reading goes in several parallel threads, which improves the reading speed. Input file is divided into set of chunks and each reading thread parses just records from this part of file.

The component can use either the fast simplistic parser (SimpleDataParser) or the robust (CharByteDataParser) one. Which parser is used depends on the component settings and data structure.

Speedup

If you use ParallelReader instead of FlatFileReader, the speed up is more significant with metadata of many date data fields.

Quoted strings

The attribute considerably changes the way your data is parsed. If it is set to true, all field delimiters inside quoted strings will be ignored (after the first Quote character is actually read). Quote characters will be removed from the field.

Example input:

1;"lastname;firstname";gender

Output with Quoted strings == true:

{1}, {lastname;firstname}, {gender}

Output with Quoted strings == false:

{1}, {"lastname}, {firstname";gender}

Examples

Reading a file with ParallelReader

This example shows the basic use of ParallelReader.

Read file file.txt using ParallelReader.

Solution

In ParallelReader, specify File URL and connect an edge to the first output port.

ParallelReader will read it using two threads.

Best Practices

We recommend users to explicitly specify Charset.

Compatibility

ParallelReader is included in CloverETL Designer version 2.8.1 and higher.

4.4.0-M1

Since 4.4.0-M1, ParallelReader support reading files from S3.

See also

FlatFileReader
Common Properties of Components
Specific Attribute Types
Common Properties of Readers
Readers Comparison