Parses data in the LibSVM Format.
Options
- delay: Delay of reading in milliseconds (default 0). DEPRECATED: use
scheduler.delay
instead if not used together with delayeach - nanodelay: Delay of reading in nanoseconds (default 0).
- delayeach: The number of lines between a delay is used (default 0).
- readfirstline: Should the first line of the file be ignored (e.g. because of header information) (default: true)
- debug: If set to true, some additional thinks are available: (default false)
- dumpEachLine: Dumps lines to the console. if set to 1 each line will be dumped
- measureEachLine: Measures the processing time between n elements that are dumped
- lastLine/maxLines: Stop processing after n elements are read
svm.delimiter: The delimiter for splitting the input (Default: ).
- svm.textDelimiter: The delimiter for strings (Default: ' ).
- addlinenumber: Adds the line number (starting with 0) to the beginning of the line. Remember to add a proper attribute to the schema! (default false).
- svm.floatingFormatter:If used for writing, each double/float value will be formatted using this formatter (default null).
- svm.numberFormatter: If used for writing, each number other than double/float value will be formatted using this formatter (default null).
- svm.trim: Removes leading and trailing whitespaces in each element (default false)
Example
PQL
CSV Protocol Handler
input = ACCESS({source='SVM', wrapper='GenericPush', transport='File',protocol='SVM', dataHandler='Tuple',options=[['delimiter',' '],['textDelimiter',"'"],['readfirstline','true'],['delay','100']], schema=[ ['symbol','String'], ['points','Double'] ] })
CQL
CSV Protocol Handler
CREATE STREAM svm (symbol String, points Double) WRAPPER 'GenericPush' PROTOCOL 'SVM' TRANSPORT 'File' DATAHANDLER 'Tuple' OPTIONS ( delimiter' ' ', 'textDelimiter' "'", 'readfirstline' 'true', 'delay' '100')