Use these instructions to build a UMF_SEARCH input document to
find entities in the entity database that match the data values of the attributes
specified in the search criteria. You send the UMF_SEARCH input document through
Web services into a Web services pipeline for processing. After the pipeline
processes the query, Web services returns a UMF_SEARCH_RESULTS output document
that contains a list of entities that matched the search critiera.
Before you begin:
The WebSphere Application Server must be running, and
at least one Web services pipeline must be started and running to receive
and process the UMF_SEARCH input document.
About this task:
Because the search request is a UMF input document, the criteria
must be formatted using valid UMF tags. You can use any text editor or utility
that creates UMF.
Procedure:
- Create a new UMF_SEARCH input document.
- In the ROOT segment, enter the required UMF tags and values, as
well as any optional UMF tags and values you want to use to specify the search
criteria. At a minimum, enter values for these UMF tags:
- Enter the data source code in the DSRC_CODE tag. The default
data source code for Web services pipeline searches is 1589. If
you use a different data source code than the default Web services pipeline
search data source code, make sure that it is configured not to resolve entities.
- Enter the data source reference code that references the requesting
message transaction in the DSRC_REF tag. The data source reference
code should be meaningful, because it is returned to the calling application.
- Enter the format code to indicate the output format of the results
using the FORMAT_CODE tag. IBM Relationship Resolution comes with
three built-in format codes for a Web services pipeline search using UMF_SEARCH:
- WS_SUMMARY_TOP10, which returns the top 10 entities that match the search
criteria
- WS_SUMMARY_TOP100, which returns the top 100 entities that match the search
criteria
- WS_SUMMARY query, which returns all entities that match the search criteria
If you use a different format code, the format code must be configured
in the UMF_OUTPUT_FORMAT table.
- Enter the minimum resolution score in the MIN_LIKE_SCORE tag
to establish the lowest numeric score that is considered a match between the
attribute values in the search criteria and the entities in the entity database
containing the same attributes. The higher the score, the more
exact the match must be. A score of 100 indicates an exact match.
- Using the other valid UMF input document segments, enter the data
values for the attributes that make up the search criteria. These
values are the attributes that the Web services pipeline search is looking
for to build the list of entities with matching or similar values. The closeness
of the match depends upon on the value in MIN_LIKE_SCORE.
- Send the UMF_SEARCH input document through Web services.
A Web services pipeline ingests the UMF_SEARCH document, using the
entity resolution process to find entities in the database using the specified
criteria. The pipeline then processes the query, creates normal logging files,
and returns the results in a UMF_SEARCH_RESULTS document through Web services
to the calling application using the selected format.
Example UMF_SEARCH document query
This example UMF_SEARCH
input document uses the WS_SUMMARY_TOP10 format code to query the entity database
to look for the top 10 entities that contain social security numbers where
the data value of the social security number exactly matches the data value
of 555-09-8761:
Note: This example is formatted for readability and does not
follow the required one line per UMF record format.
<UMF_SEARCH>
<DSRC_CODE>1589</DSRC_CODE>
<DSRC_REF>1223</DSRC_REF>
<MIN_LIKE_SCORE>100</MIN_LIKE_SCORE>
<FORMAT_CODE>WS_SUMMARY_TOP10</FORMAT_CODE>
<NUMBER>
<NUM_TYPE>SSN</NUM_TYPE>
<NUM_VALUE>555-09-8761</NUM_VALUE>
</NUMBER>
</UMF_SEARCH>