Detail on the WITSML to PPDM mapping project, a joint initiative betwwen the PPDM Association and Energistics to standardise movement of E&P data in the Oil & Gas industry. We outline the project and place it in the context of a data management approach to E&P data.
2. Objectives:
• Standard baseline
mapping to/from
PPDM
• Determine if there is
value to mapping all
WITSML data objects
• Define change
management process
for projects involving
2+ standards
organizations
WITSML-PPDM Mapping Project
The WITSML-PPDM Mapping Project is a joint
Energistics-PPDM initiative in the Oil & Gas industry
6. Resolving identification
• Determine whether an incoming well already exists
• Tie trajectories to the right wellbore in PPDM
• Use alias table, but some of this is implementation
defined, and it must be correct or you can end up with
all kinds of data quality issues
7. What makes an implementation valid?
Some examples:
• Well formed
• Schema validity
• Standards validity
• Application validity
The Project’s views on mapping validity
8. Valid XML
Valid SQL queries
Achieved by good development practices
Well formed
9. Automated checks by industry standard tools
WITSML schema
• Ordering –
sequence/alternate
groups
• Cardinality
• Mandatory
Elements/Attributes
PPDM
• Constraints:
PK, FK, NOT
NULL, Check
Schema validity
11. Can your applications read what you have written?
Fail loudly
• Queries fail
Fail quietly
• Missing data
• Spurious data
• Wrong data
Application validity
12. • Relational systems
• XML systems
• Other systems
• Mappings between
systems
Adopting a data management approach to the Project
PPDM data management module:
21. WITSML example files
Example files help us to understand what is required.
All the identifiers here have rich definitions in the schema.
22. Structure and ordering
The schema will tell you about the structure of the
WITSML document; not just the parent/child hierarchy but
the order of elements
23. Test cases
The schema can also help with test cases. This is a very
small, valid WITSML well (some XML namespace
information is missing).
24. PPDM data model
There is similar machine-readable information on PPDM:
data types, constraints, documentation, null values etc
25. • Understand implications
• Manually regurgitate into
implementation
Implementation benefits: We can read it
26. • Analyse
Check for potential issues
• Generate
Implementation (with help)
Auditing
• Report
Timely, accurate
information on data flow
Implementation benefits: The computer can read it
27. • Data movement is part of
data management
• A sustainable strategy is
required
• Much work is inadvertently
duplicated
• We can reduce the effort
and improve our ability to
maintain data mappings
Conclusion
28. Watch an E&P data management
demo by Richard Cook, our senior
E&P specialist
http://www.etlsolutions.com/what-we-do/oil-and-gas/
Read our free white paper on
PPDM data integration:
http://www.etlsolutions.com/wp-
content/uploads/2012/07/Whitepaper-on-PPDM-Data-
Integration-Sept12.pdf
Free resources
Images via http://www.freedigitalphotos.net
Hinweis der Redaktion
Objectives
standard baseline mapping to/from PPDM
determination if there is value to mapping all WITSML data objects
define change management process for projects involving 2+ standards organizations
Deliverables
spreadsheet, PPDM mapping table, use cases, documentation (note: pilot includes a sub-set of WITSML data-objects and associated PPDM tables that support selected use cases)
Target completion for pilot: Q4 2013 (publish Q1 2014)
This slide reminds me to talk about identification. More details in the paper. Incoming WITSML has uids, locally unique. How to match to PPDM. Need to see if an incoming well already exists, tie trajectorys to the right wellbore in PPDM etc.
Use alias table, but some of this is implementation defined, and it needs to be correct or you can end up with all kinds of data quality issues. Sometimes this is almost irrelevant – incoming well uid is the same as the UWI for example.
This slide reminds me to talk about identification. More details in the paper. Incoming WITSML has uids, locally unique. How to match to PPDM. Need to see if an incoming well already exists, tie trajectorys to the right wellbore in PPDM etc.
Use alias table, but some of this is implementation defined, and it needs to be correct or you can end up with all kinds of data quality issues. Sometimes this is almost irrelevant – incoming well uid is the same as the UWI for example.
What do we mean by a data mgmt. approach to mapping, and what are the implications?
What do we mean by a data mgmt. approach to mapping, and what are the implications?
What do we mean by a data mgmt. approach to mapping, and what are the implications?
What do we mean by a data mgmt. approach to mapping, and what are the implications?
What do we mean by a data mgmt. approach to mapping, and what are the implications?
What do we mean by a data mgmt. approach to mapping, and what are the implications?
PPDM allows us to describe systems and mapping between systems. We immediately gain the benefits of referential integrity. The mapping references tables, columns and schema entities, and ref integ ensures that they exist in the model. We also get a wealth of additional information about individual mappings: documentation, column types, constraints, etc.
We can also query the mappings, generate different views of them, documentation etc. and it’s easy to do “what-ifs” – what if we upgrade to PPDM 3.9.
Example files help us to understand what is required. But note that all the identifiers here – name, nameLegal etc, have rich definitions in the schema.
The schema will tell you about the structure of the WITSML document, not just the parent/child hierarchy, but the order of elements.
The schema can also help us with test cases. Here’s a very small valid WITSML well (some XML namespace information is missing). What happens if you get invalid XML
We have similar machine readable information on PPDM: data types, constraints (FK/PK), documentation, null, etc.
Existence of spec doesn’t mean you can just hand to developers and expect it to meet your business needs. Data provenance may mean you use the PPDM RM and data management modules to record the original data, how it was processed, etc. How you intend to use the data changes how/where in PPDM you store it, eg just metadata, values stored in BLOB or file ref’d by RM, or in WLC_VALUE to give consistent interface, ability to reconsititute data in a number of different ways (eg get all logs for a WELL in LAS format).
Existence of spec doesn’t mean you can just hand to developers and expect it to meet your business needs. Data provenance may mean you use the PPDM RM and data management modules to record the original data, how it was processed, etc. How you intend to use the data changes how/where in PPDM you store it, eg just metadata, values stored in BLOB or file ref’d by RM, or in WLC_VALUE to give consistent interface, ability to reconsititute data in a number of different ways (eg get all logs for a WELL in LAS format).