Package edu.caltech.nanodb.storage

Interface TupleFileManager

  • All Known Implementing Classes:
    HeapTupleFileManager
    public interface TupleFileManager
    This interface defines the operations that can be performed on TupleFiles, but that are at a higher level of implementation than the tuple file itself. Examples of such operations are creating a new tuple file on disk, deleting a tuple file from the disk, and so forth.

    • Method Detail

      • getShortName

        java.lang.String getShortName()
        Returns the "short name" used to refer to this storage-file type. This is the name used in e.g. CREATE TABLE commands to specify that a table-file should use a particular storage-file type.
        Returns:
        the "short name" used to refer to this storage-file type.

      • getDBFileType

        DBFileType getDBFileType()
        Returns the DBFileType value used to indicate this storage format in data files.
        Returns:
        the DBFileType value used to indicate this storage format in data files.

      • createTupleFile

        TupleFile createTupleFile​(DBFile dbFile,
                          Schema schema)
        Initialize the specified DBFile to be a new empty table with the specified schema.
        Parameters:
        dbFile - the file to use for the new tuple-file
        schema - the schema of the table data to store in the file
        Returns:
        a TupleFile object that allows tuple operations to be performed against the DBFile.

      • openTupleFile

        TupleFile openTupleFile​(DBFile dbFile)
        Open the specified DBFile as a TupleFile of a specific type, containing tuples of data conforming to a specific schema.
        Parameters:
        dbFile - the tuple-file to open
        Returns:
        a TupleFile object that allows tuple operations to be performed against the DBFile.

      • saveMetadata

        void saveMetadata​(TupleFile tupleFile)
        Writes the metadata (schema and stats information) for the tuple-file back into the tuple-file. The schema and statistics are represented as separate objects (Schema and TableStats objects, specifically), so of course they need to be serialized into a binary representation for storage when they change.
        Parameters:
        tupleFile - the tuple-file whose metadata should be persisted

      • deleteTupleFile

        void deleteTupleFile​(TupleFile tupleFile)
        Delete the specified tuple-file.
        Parameters:
        tupleFile - the tuple-file to delete