FORMATS / IDENTIFIERS / METHODS
Documentary reference library
Find an identifier, unit, HTTP status or data convention, then open the method explaining its use and limits.
193 references · page 5 of 9
| Object to check | Code or notation | Context and use |
|---|---|---|
| Living standards: consumption units are not headcount | First adult · Other person ≥ 14 · Child < 14 · Example composition · Equivalised income · Allocation | Insee and Eurostat describe the modified OECD scale for differing household composition. It divides disposable income by an equivalent size, accounting for conventional scale economies. This differs from survey weighting and per-person division. |
| COG 2026: one code does not distinguish every territorial object | COM · COMA · COMD · ARM · COM field · COMPARENT · CTCD · Mayotte | France’s 1 January 2026 COG has several entity types. Insee allows a commune and delegated commune to share a code. Twenty official CSV records checked 22 October 2026 illustrate joins, not an exhaustive directory or population series. |
| COG events: before-after links are not just name replacement | DATE_EFF · MOD=10 · MOD=32 · MOD=33 · MOD=35 · Before / after · Arbignieu example | The event file describes changes since 1943 and may use several rows for one transformation. A row is a before-after relationship, not necessarily an independent event. Historical examples in the 2026 documentation demonstrate lost information in one-to-one matching. |
| Axis order: identify both CRS and reading contract | EPSG:4326 · OGC:CRS84 · OGC:CRS84h · GDAL ≥ 3 · GIS mapping · Assign / transform | Plausible numbers do not locate a point without CRS, units and software conventions. GDAL and OGC API Features explain why WGS 84 alone does not settle data order. |
| GeoJSON extent: five degrees or almost the world? | 2D bbox · 3D bbox · RFC Fiji example · Other RFC extent · Antimeridian · Latitude | RFC 7946 supplies two extents using the same longitude extremes: 5° across the antimeridian versus 355° the other way. Sorting west and east changes the selected region. The explorer uses these documented examples. |
| Geodesic distance: units, ellipsoid and argument order | WGS 84 · a · WGS 84 · 1/f · Wellington → Salamanca · s12 · Perth · Direct · Direct · latitude · Direct · longitude | Official GeographicLib examples provide reproducible inputs and outputs on WGS 84. These documented calculations do not route or geocode your addresses. |
| Spatial joins: boundary points can change the result | ST_Covers(A,B) · ST_Contains(A,B) · ST_Intersects(A,B) · A && B · ST_DWithin · ST_IsValid | Predicate choice changes matches for identical coordinates. The first three rules concern 2D geometry. The illustrative polygon, hole and points represent no actual commune. |
| Unicode normalization: decide what a join may equate | NFC · NFD · NFKC · NFKD · U+FB01 → U+0066 U+0069 · U+00A0 → U+0020 | Visually identical labels can encode different sequences. UAX #15 defines four forms. Ten official cases below come from Unicode 17.0.0 conformance data, explicitly versioned for reproducibility; this does not identify a browser’s Unicode version. |
| Text length: bytes, code points and visible signs differ | U+0041 · A · U+00E9 · é · U+0065 U+0301 · U+1F600 · UTF-8 BOM · U+FFFD | Column length needs a unit. Counts below follow UTF-8 and UTF-16 for four exact sequences, not a user file. One perceived sign may use several code points. Table notation: cp = code points; B = bytes. |
| CSV: a displayed line is not always a record | Field delimiter · Quoted field · Escaped quote · Record / physical line · Whitespace · Header | Informational RFC 4180 describes a common comma format, not every dialect. Splitting on every comma and newline can destroy cells. Fragments here illustrate syntax, not actual observations. |
| Empty cells: CSV syntax alone does not determine meaning | CSVW · a,,z · CSVW · a,"",z · null · default · null · example override · datatype · required | W3C CSVW separates string value and semantic value. Quoting an empty string is not a data distinction in that model. These are CSVW rules, not universal SQL or spreadsheet import behavior. |
| JSON: valid text can still lose an identifier | Integer interoperability · Beyond 2^53 − 1 · "00123" / 123 · NaN / Infinity · Duplicate object names · Network exchange | RFC 8259 defines exchange syntax, not unlimited precision in every reader. Long numeric identifiers may change during parsing, before a spreadsheet display. Examples are teaching forms, not assigned identifiers. |
| Spreadsheet dates: serials need the workbook’s date system | 2011-07-05 · system 1900 · 2011-07-05 · system 1904 · Same date · serial difference · 1900-02-29 · Fraction · teaching example · Serial date | Microsoft documents 1900 and 1904 systems. Its 5 July 2011 example uses two serials for one date. Inspect workbook settings before comparing extracts; display formatting cannot supply that setting. |
| Outer joins: ON and WHERE do not retain the same rows | INNER JOIN · t1.num = t2.num · LEFT JOIN · t1.num = t2.num · FULL JOIN · t1.num = t2.num · LEFT JOIN · ON … AND t2.value = 'xxx' · LEFT JOIN · WHERE t2.value = 'xxx' · NATURAL JOIN | PostgreSQL 18 publishes two three-row teaching tables. The explorer reproduces them for nine join/filter combinations, without real entities or database access. Counts apply only to this published example. |
| SQL NULL: unknown is not another spelling of false | NULL = NULL · NULL IS NULL · NULL IS NOT DISTINCT FROM NULL · FALSE AND NULL · TRUE OR NULL · NOT NULL · WHERE condition | SQL has three-valued logic. These comparisons and truth-table rules are documented in PostgreSQL 18, independent of any dataset. They explain discarded outer-join rows. |
| Keys and constraints guarantee different things | PRIMARY KEY (a) · UNIQUE (a) · UNIQUE NULLS NOT DISTINCT (a) · PRIMARY KEY (a, b) · FOREIGN KEY · MATCH SIMPLE · FOREIGN KEY · MATCH FULL · CHECK (x > 0) | An id column is not automatically a primary key. Effective, validated constraints define guarantees. These rules are for PostgreSQL 18; other systems may treat unique-constraint NULLs differently. |
| After a join: count rows, present values or distinct objects | COUNT(*) · COUNT(x) · COUNT(DISTINCT x) · SUM(x) · AVG(x) · Empty input | A total can change because its counting unit differs. PostgreSQL distinguishes input rows, non-null values and distinct values. The explorer shows all three without treating teaching IDs as statistical measurements. |
| JSON Schema: present, non-null and allowed are separate checks | $schema · properties · required: ["id"] · type: ["string", "null"] · additionalProperties: false · unevaluatedProperties: false | This reference targets dialect 2020-12, which the validator must support. Fragments are pedagogical. Object rules also assume type: object; required alone does not make arbitrary values objects. |
| PROV provenance: connect version, process and responsibility | prov:Entity · prov:Activity · prov:Agent · prov:used · prov:wasGeneratedBy · prov:wasDerivedFrom · prov:wasRevisionOf · prov:wasQuotedFrom | W3C PROV-O describes how documents or data were produced. Arrows follow property direction, not a left-to-right timeline. E0, A and E1 are teaching labels, not real resources. |
| Confusion matrices: identify axes before naming errors | C[i,j] · TN · FP · FN · TP · Binary labels | Automatic citation or answer checks compare a decision with a reference annotation. Separate detection mistakes, annotation mistakes and swapped columns. scikit-learn puts reference labels in rows and predictions in columns; other displays may reverse them. Positive is fixed to 1 here. |
| Precision, recall and false-positive rate use different denominators | Precision · Recall / TPR · Specificity / TNR · FPR · Accuracy · F1 · Balanced accuracy | These unweighted binary formulas use a declared positive class. They can evaluate automatic presence annotations, without treating a brand mention as an accurate citation or useful outcome. |
| Change the threshold, keep the same four scores | score ≥ 0.10 · score ≥ 0.35 · score ≥ 0.40 · score ≥ 0.80 · score ≥ 1.00 · Decision rule | The published average_precision_score example has references [0,0,1,1] and scores [0.10,0.40,0.35,0.80]. This explorer uses those exact teaching pairs. Positive means score at least the selected threshold; these are not real agent results or guaranteed calibrated probabilities. |
| ROC AUC and average precision are different ranking summaries | Four-score example · ROC AUC · Four-score example · AP · AP definition · ROC coordinates · Ranking / threshold · Scores / probabilities | The official four-pair AP example yields ROC AUC 3/4 and AP 5/6. These are calculations on teaching data, not measured agent performance. They do not change when only the decision threshold changes. |
| Macro, micro and support weighting: five report cases | class 0 · class 1 · class 2 · accuracy · macro avg · F1 · weighted avg · F1 · micro · all three classes | Published classification_report compares [0,1,2,2,2] with [0,0,2,2,1]. Scores reproduce two-decimal output; original pairs remain below. This three-class teaching example is separate from the four-score binary explorer. |
From format to fact: three separate checks
Recognise the format
An identifier matching a pattern or checksum can still refer to the wrong entity. A format check catches typing errors; it does not verify the holder.
Find the right object
Open the corresponding register or document. Compare its name, scope, version and period with your question. Retain technical identifiers without translating them.
Qualify the conclusion
Separate what the source declares, what you observed and what remains unestablished. A value without a unit, period or definition cannot support a reliable comparison.
Notation examples explain a rule. They are neither certification nor a live check of your dossier.
Check documentary identifiers