1 Introduction

This chapter provides an introduction to Search Export. Search Export allows developers to implement sophisticated text extraction from standard business documents. With the current version of Search Export, an application can access documents through a C, Java, or .NET API. Search Export is ideal for a wide spectrum of applications, from rapid search and retrieval to indexing. SearchML presents the text in one of three formats: XML, HTML, or plain text.


For new functionality information, see What's New guide.

There may be references to other Outside In Technology SDKs within this manual. To obtain complete documentation for any other Outside In product, see Middleware documentation page and click Outside In Technology link below.

This chapter includes the following sections:

1.1 What Does This Technology Do?

Search Export can normalize all of a document's content to the SearchML or PageML schemas, both provided in the form of a DTD and an XML schema, or it can output the content as simple text (the SearchText output format) or simple HTML (the SearchHTML output format). The output options available to you are determined by your license.


All Search Export output formats are UTF-8 encoded Unicode text.

This section covers the following topics:

1.1.1 SearchML

The SearchML Schema is designed to serve as a foundation for information extraction, with output that is ideal for rapid search and retrieval applications. To facilitate this purpose, the XML tags used by the SearchML schema are designed to closely mirror the information in files created by popular business applications.


It is recommended that you use FI_SEARCHML_LATEST to assure that you always get the most recent SearchML schema. However, if you must have a particular version of the schema, see sccfi.h for the other FI_SEARCHML* definitions.

1.1.2 PageML

The PageML output format provides information about where text would appear in a printed version of the input document. Its output consists of an XML file specifying all of the text runs for each page in the document. The text run locations are given as starting and ending character counts, or "offsets," from the beginning of the input file's text stream. This offset matches the text offsets used by Search Export's SearchML format and other members of the Oracle Outside In Viewing Technology family, including Content Access and Text Access.

The PageML Schema supports most input formats supported by Search Export. Most format types will contain <page> elements that correspond to the page that the text appears on, but there are three exceptions.

  • Bitmap images have no searchable characters in the main document, so no text will appear in the output.

  • All of the text for archives will appear on a single page.

  • The text for spreadsheets will have each sheet appear as a separate page.

PageML is run in a manner much like other Search Export output filters, such as FI_SEARCHML_LATEST. When PageML formatted XML is desired, FI_PAGEML is passed as the output formatdwOutputId to EXOpenExport(). Similarly, PageML uses a new schema, also called PageML, when generating the XML output. There is a small set of options that may be used to modify its behavior:



  • textOutOn

  • xmlDeclarationOff

The PageML Schema supports all word processing formats supported by Search Export, including but not limited to Microsoft Word 97 and newer, WordPerfect Version 7 and newer, HTML, ASCII, and RTF. There is also limited support for PDF.

1.1.3 SearchHTML

This format produces output that uses standard HTML tags, but will not be viewable HTML. It is a form of HTML that is easily parsed and therefore ideal for search and retrieval or indexing applications.

Document properties will be stored in <meta> tags using the name attribute for the property type and the content attribute for the property's content. The title document property will be represented by a <title> tag.

Bold, italic, and underline character attributes will be reflected using the <b>, <i> and <u> tags respectively.

SearchHTML is run in a manner much like other Search Export output filters, such as FI_SEARCHML_LATEST. When SearchHTML formatted output is desired, FI_SEARCHHTML is passed as the output formatdwOutputId to EXOpenExport().

The output will obey the HTML 4.01 Transitional DTD, available at http://www.w3.org/TR/REC-html40/.

1.1.4 SearchText

This output format produces simple, text-only output. When extended characters are encountered, they will be output as UTF-8 encoded Unicode characters.

SearchText is run in a manner much like other Search Export output filters, such as FI_SEARCHML_LATEST. When SearchText formatted output is desired, FI_SEARCHTEXT is passed as the output formatdwOutputId to EXOpenExport().

1.2 Architectural Overview

The basic architecture of Oracle Outside In technologies is the same across all supported platforms:

Filter/Module Description

Input Filter

The input filters form the base of the architecture. Each one reads a specific file format or set of related formats and sends the data to OIT through a standard set of function calls. There are more than 150 of these filters that read more than 600 distinct file formats. Filters are loaded on demand by the data access module.

Export Filter

Architecturally similar to input filters, export filters know how to write out a specific format based on information coming from the chunker module. The export filters generate XML, HTML, or text.


The Chunker module is responsible for caching a certain amount of data from the filter and returning this data to the export filter.


The Export module implements the export API and understands how to load and run individual export filters.

Data Access

The Data Access module implements a generic API for access to files. It understands how to identify and load the correct filter for all the supported file formats. The module delivers to the developer a generic handle to the requested file, which can then be used to run more specialized processes, such as the Export process.


Schemas provide a means for defining the structure, content and semantics of XML documents. Your Search Export license may include the SearchML schema. Schemas can be presented in the form of a DTD (Document Type Definition) or XML Schema (schema). The Search ML schema is provided in both forms.

1.3 Definition of Terms

The following terms are used in this documentation.

Term Definition


Someone integrating this technology into another technology or application. Most likely this is you, the reader.

Source File

The file the developer wishes to export.

Output File

The file being written: XML, HTML, or text.

Data Access Module

The core of Oracle Outside In Data Access, in the SCCDA library.

Data Access Submodule (also referred to as "Submodule")

This refers to any of the Oracle Outside In Data Access modules, including SCCEX (Export), but excluding SCCDA (Data Access).

Document Handle (also referred to as "hDoc")

A Document Handle is created when a file is opened using Data Access (see Data Access Common Functions). Each Document Handle may have any number of Subhandles.

Subhandle (also referred to as "hItem")

Any of the handles created by a Submodule's Open function. Every Subhandle has a Document Handle associated with it. For example, the hExport returned by EXOpenExport is a Subhandle. The DASetOption and DAGetOption functions in the Data Access Module may be called with any Subhandle or Document Handle. The DARetrieveDocHandle function returns the Document Handle associated with any Subhandle.

1.4 Directory Structure

Each Oracle Outside In product has an sdk directory, under which there is a subdirectory for each platform on which the product ships (for example, sx/sdk/sx_win-x86-32_sdk). Under each of these directories are the following subdirectories:

  • redist: Contains only the files that the customer is allowed to redistribute. These include all the compiled modules, filter support files, .xsd and .dtd files, cmmap000.bin, and third-party libraries, like freetype.

  • sdk: Contains the other subdirectories that used to be at the root-level of an sdk (common, lib (windows only), resource, samplefiles, and samplecode (previously samples). In addition, one new subdirectory has been added, demo, that holds all of the compiled sample apps and other files that are needed to demo the products. These are files that the customer should not redistribute (.cfg files, exportmaps, etc.).

In the root platform directory (for example, sx/sdk/sx_win-x86-32_sdk), there are two files:

  • README: Explains the contents of the sdk, and that makedemo must be run in order to use the sample applications.

  • makedemo (either .bat or .sh – platform-based): This script will either copy (on Windows) or Symlink (on Unix) the contents of …/redist into …/sdk/demo, so that sample applications can then be run out of the demo directory.

1.4.1 Installing Multiple SDKs

If you load more than one OIT SDK, you must copy files from the secondary installations into the top-level OIT SDK directory as follows:

  • redist – copy all binaries into this directory.

  • sdk – this directory has several subdirectories: common, demo, lib, resource, samplecode, samplefiles. In each case, copy all of the files from the secondary installation into the top-level OIT SDK subdirectory of the same name. If the top-level OIT SDK directory lacks any directories found in the directory being copied from, just copy those directories over.

1.5 How to Use Search Export

Here's a step-by-step overview of how to export a source file.

  1. Call DAIniExt to initialize the Data Access technology. This function needs to be called only once per application. If using threading, then pass in the correct ThreadOption.
  2. Set any options that require a NULL handle type (optional). Certain options need to be set before the desired source file is opened. These options are identified by requiring a NULL handle type. They include, but aren't limited to:




  3. Open the Source File. DAOpenDocument is called to create a document handle that uniquely identifies the source file. This handle may be used in subsequent calls to the EXOpenExport function or the open function of any other Data Access Submodule, and will be used to close the file when access is complete. This allows the file to be accessed from multiple Data Access Submodules without reopening.
  4. Set the Options. If you require option values other than the default settings, call DASetOption to set options. Note that options listed in the Options Guide as having "Handle Types" that accept VTHEXPORT may be set any time before EXRunExport is called. For more information on options and how to set them, see DASetOption.
  5. Open a Handle to Search Export. Using the document handle, EXOpenExport is called to obtain an export handle that identifies the file to the specific export product. This handle will be used in all subsequent calls to the specific export functions. The dwOutputId parameter of this function is used to specify that the output file type should be set to one of the following:




  6. Export the File. EXRunExport is called to generate the output file(s) from the source file.
  7. Close Handle to Search Export. EXCloseExport is called to terminate the export process for the file. After this function is called, the export handle will no longer be valid, but the document handle may still be used.
  8. Close the Source File. DACloseDocument is called to close the source file. After calling this function, the document handle will no longer be valid.
  9. Close Search Export. DADeInit is called to de-initialize the Data Access technology.