TextLine reads and writes raw text files and returns tuples which, by default, contain two fields specific to the platform used. The first field is either the byte offset or line number, and the second field is the actual line of text. When written to, all Tuple values are converted to Strings delimited with the TAB character (\t). A TextLine scheme is provided for both the local and Hadoop modes.

By default TextLine uses the UTF-8 character set. This can be overridden on the appropriate TextLine constructor.

TextDelimited reads and writes character-delimited files in standard formats such as CSV (comma-separated variables), TSV (tab-separated variables), and so on. When written to, all Tuple values are converted to Strings and joined with the specified character delimiter. This Scheme can optionally handle quoted values with custom quote characters. Further, TextDelimited can coerce each value to a primitive type when reading a text file. A TextDelimited scheme is provided for both the local and Hadoop modes.

By default TextDelimited uses the UTF-8 character set. This can be overridden on appropriate the TextDelimited constructor.

SequenceFile is based on the Hadoop Sequence file, which is a binary format. When written to or read from, all Tuple values are saved in their native binary form. This is the most efficient file format - but be aware that the resulting files are binary and can only be read by Hadoop applications running on the Hadoop platform.

Like the SequenceFile Scheme, WritableSequenceFile is based on the Hadoop Sequence file, but it was designed to read and write key and/or value Hadoop Writable objects directly. This is very useful if you have sequence files created by other applications. During writing (sinking), specified key and/or value fields are serialized directly into the sequence file. During reading (sourcing), the key and/or value objects are deserialized and wrapped in a Cascading Tuple object and passed to the downstream pipe assembly. This class is only available when running on the Hadoop platform.

The cascading.tap.local.FileTap tap is used with the Cascading local platform to access files on the local file system.

The cascading.tap.hadoop.Hfs tap uses the current Hadoop default file system, when running on the Hadoop platform.

If Hadoop is configured for "Hadoop local mode" (not to be confused with Cascading local mode), its default file system is the local file system. If configured for distributed mode, its default file system is typically the Hadoop distributed file system.

Note that Hadoop can be forced to use an external file system by specifying a prefix to the URL passed into a new Hfs tap. For instance, using "s3://somebucket/path" tells Hadoop to use the S3 FileSystem implementation to access files in an Amazon S3 bucket. More information on this can be found in the Javadoc.

The cascading.tap.MultiSourceTap is used to tie multiple tap instances into a single tap for use as an input source. The only restriction is that all the tap instances passed to a new MultiSourceTap share the same Scheme classes (not necessarily the same Scheme instance).

The cascading.tap.MultiSinkTap is used to tie multiple tap instances into a single tap for use as output sinks. At runtime, for every Tuple output by the pipe assembly, each child tap to the MultiSinkTap will sink the Tuple.

The cascading.tap.hadoop.TemplateTap and cascading.tap.local.TemplateTap are used to sink tuples into directory paths based on the values in the Tuple. More can be read below in Template Taps.

The cascading.tap.hadoop.GlobHfs tap accepts Hadoop style "file globbing" expression patterns. This allows for multiple paths to be used as a single source, where all paths match the given pattern. This tap is only available when running on the Hadoop platform.