An operator takes a position in Dutch acreage and receives the regulator's public log package with it: seven wells logged between 1958 and 1990, in LAS 2.0, arriving in two tranches. The logs are real NLOG composite logs as the OSDU Forum redistributes them, 4376 depth steps in total. One external table spans the whole library even though the curve sets differ by thirty years of tool development, because the reader merges the per-file schemas into a union: gamma ray on all seven wells, sonic and density on three, and one 1990 well carrying nothing but a gamma ray. A well whose tool never recorded a sonic gets NULL for it, which is the truthful answer and not a zero. What this demo shows that the DLIS and LIS ones cannot is that the LAS well-information section rides on every row: the library is queryable by well, field and vintage with no join to a header table anywhere, and the well name comes out of the file's own header rather than its name. The load also turns the literal string UNKNOWN, which NLOG writes for an unattributed field, into a real NULL so it stops grouping as though it were a field. The loader keys on the source file, so one tranche is run twice and the count asserted unchanged.
-- ============================================================================
-- Well Log Library Consolidation - Setup Script
-- ============================================================================
-- An operator takes a position in Dutch acreage and receives the regulator's
-- public log package with it. Seven wells logged between 1958 and 1990, in
-- LAS 2.0, arriving in two tranches.
--
-- The logs are real. They are NLOG composite logs from the Netherlands, as
-- the OSDU Forum redistributes them; see ATTRIBUTION.md in the parent folder.
-- Their curve sets differ by vintage because the tools did: the three
-- Wassenaar wells from 1958 to 1963 carry a gamma ray and a neutron and
-- nothing else, the wells from 1969 on carry the full triple combo, and one
-- 1990 well carries only a gamma ray.
--
-- 11 March MED-01 (1958), WAS-25 (1961), MED-05 (1963) 1585 depth steps
-- 12 March D15-01 (1969), K08-02 (1972), GRW-01 (1979),
-- L09-06 (1990) 2791 depth steps
--
-- One external table spans the whole library even though the curve sets
-- differ, because the reader merges the per-file schemas into a union: a well
-- whose tool never recorded a sonic gets NULL for it, which is the truthful
-- answer and not a zero.
--
-- 1. log_files external, DISCOVER over the landing folder
-- 2. log_library DELTA, the curated library
-- ============================================================================
-- ----------------------------------------------------------------------------
-- STEP 1: Zone and schema
-- ----------------------------------------------------------------------------
CREATE ZONE IF NOT EXISTS {{zone_name}} TYPE EXTERNAL
COMMENT 'External tables - demo datasets and file-backed data';
CREATE SCHEMA IF NOT EXISTS {{zone_name}}.log_library
COMMENT 'NLOG well log package read in place from LAS, curated into Delta';
-- ----------------------------------------------------------------------------
-- STEP 2: Register the landing folder with DISCOVER
-- ----------------------------------------------------------------------------
-- LAS is sectioned text, identified by its `~V` and `~C` section markers
-- rather than by a magic number.
--
-- The well-information section rides on every row as constant columns:
-- well_well, well_fld, well_strt and the rest. That is the whole reason a
-- table over a log library is queryable by well or field with no join to a
-- header table, and it is why the curated table below takes the well name
-- from well_well rather than from the file name. A file renamed in transit
-- still says which well it holds.
-- ----------------------------------------------------------------------------
DROP EXTERNAL TABLE IF EXISTS {{zone_name}}.log_library.log_files;
DISCOVER {{zone_name}}.log_library.log_files
PATH '{{data_subdir}}/landing'
WITH (FILE_METADATA = true);
-- ----------------------------------------------------------------------------
-- STEP 3: The curated Delta library
-- ----------------------------------------------------------------------------
-- Every curve is nullable because the library genuinely does not have them
-- all. Four of the seven wells were logged before a compensated density tool
-- was on the truck, and a curated table that filled those with zero would
-- make a 1958 well look like a rock with no density rather than a well that
-- was never measured for one.
-- ----------------------------------------------------------------------------
CREATE DELTA TABLE IF NOT EXISTS {{zone_name}}.log_library.log_library (
well VARCHAR,
field VARCHAR,
vintage INTEGER,
delivered_on VARCHAR,
source_file VARCHAR,
dept DOUBLE,
gr DOUBLE,
dt DOUBLE,
rhob DOUBLE,
drho DOUBLE,
nphi DOUBLE
) LOCATION '{{data_subdir}}/curated/log_library';## What this demo does An operator takes a position in Dutch acreage and receives the regulator's public log package with it: seven wells logged between 1958 and 1990, in LAS 2.0, arriving in two tranches. The logs are real NLOG composite logs as the OSDU Forum redistributes them, 4376 depth steps in total. One external table spans the whole library even though the curve sets differ by thirty years of tool development, because the reader merges the per-file schemas into a union: gamma ray on all seven wells, sonic and density on three, and one 1990 well carrying nothing but a gamma ray. A well whose tool never recorded a sonic gets NULL for it, which is the truthful answer and not a zero. What this demo shows that the DLIS and LIS ones cannot is that the LAS well-information section rides on every row: the library is queryable by well, field and vintage with no join to a header table anywhere, and the well name comes out of the file's own header rather than its name. The load also turns the literal string UNKNOWN, which NLOG writes for an unattributed field, into a real NULL so it stops grouping as though it were a field. The loader keys on the source file, so one tranche is run twice and the count asserted unchanged. ## What it creates - `{{zone_name}}.log_library.log_files` (external table) - `{{zone_name}}.log_library.log_library` (delta table) ## Running it The demo ships `setup.sql`, `queries.sql` and `cleanup.sql` under `demos/subsurface/las-well-log-library`. The queries carry their own `ASSERT`s, so a run that completes has verified its own numbers. Data: 7 files, 0.2 MB.