Pool raw plug values across wells in one call

You need a single long-form DataFrame of plug values + grouping columns, pooled across every well, so you can compute pooled percentiles, build diagnostic plots, or merge with external metadata. No per-well loop.

The minimal pattern

df = manager.PHIE.filter("Facies").data(weighted=True)

Result columns:

well	DEPT	PHIE	Facies	Weight
Well_A	2500.0	0.193	Clean	0.25
Well_A	2500.5	0.062	Medium	0.50
…	…	…	…	…

• well — original well name (not the sanitized key).

• DEPT — depth.

• PHIE — the property values; uses your label strings if you set Property.labels.

• Facies — one column per active filter, named after the filter property (no Group1/Group2).

• Weight — depth-interval weight per row, present when weighted=True. Use it for depth-weighted percentiles externally: np.average(df["PHIE"], weights=df["Weight"]).

Wells lacking the property are silently skipped (with a warning). Pass warn_missing=False to silence the warning in scripted use.

Multiple properties at once

df = manager.properties(["PHIE", "PERM"]).filter("Facies").data()

Yields one column per property plus the filter columns; depths are joined per well so the result is one row per (well, depth).

Built-in summary

The same long-form data is available aggregated:

manager.PHIE.filter("Facies").stats(
    return_df=True,
    flat_columns=True,           # Facies column instead of "Group"
    methods=["mean", "percentile_50", "percentile_90"],
)

Verifying

• story_tests/story_3_pooled_data.py shows the basic shape.

• story_tests/story_6_pooled_extraction.py walks three real use cases — pooled percentile table, suspect-plug diagnostic subset, and cross-check against the library aggregation.