Reference Manual: DSL textParser

This is a utility class that implements a simple parser for text files. The current implementation is not necessarily efficient and, therefore, for large text files we encourage a specialized user implementation.

Info

This method has been removed from the API see [SMILE/SMILearn (C++): http://genie.sis.pitt.edu/download/smile_api_changes.txt change notes] Use instead :

  DSL_dataset dataset;
  std::string filename;

  if (!dataset.ReadFile(filename))   _error_message();

Methods

• DSL_textParserParams

DSL_textParserParams is a structure that stores settings for the DSL_textParser. It has the following fields and their default values:

• bool useHeader = true, defines if the first row in the text file contains the labels of the variables.

• bool createPreprocessFile = false, when set to true, the parser produces a text file with parsed information. The file carries the name of original data file and extension .glw.

• std::string textFileName = “”, name of a text file that contains the data.

• std::string missingValue = “*”, the marker in the data that corresponds to a missing element.

• bool typesSpecified = false, the marker that indicates if in the header of the data types are specified. There are three possible keywords: ordinal, discrete, and continuous. By default, the parser assumes the discrete type (values are treated as labels).

• DSL_textParser()

DSL_textParser(const DSL_textParserParams &params)

The constructors for the file parser. The first constructor uses the default values of the DSL_textParserParams, while the second allows the user to specify values explicitly.

• int Parse();

This method performs actual file parsing and creates internal data structures. Every time this method is called, the file parsing takes place. It returns DSL_OKAY if the operation was successful, and an error code in case of an error.

• void GetDataset(DSL_dataset &here)

This method allows obtaining a data set from parsed file. The method Parse() should be called prior to this method.