Read one table. Call it before writing SQL against a table, with purpose: query: DDL, the Joins list (declared foreign keys, joins dbprint inferred from column names or measured from values, and joins a person added, tagged annotated - minus any a person rejected; a join mentioned only in the table's prose description is not in it), a data dictionary, the value lists a predicate is written from, with their counts and coverage, and each nullable column's null share, plus notes that apply to the whole database. Call it with the default purpose: profile to describe the data - statistics such as null rates, cardinality and ranges, joins, the table's prose description and notes. Use search_columns or list_tables first when the table is not yet known, and resolve_value to check how one phrase is spelled in one column. Paged: md and yaml end every page but the last with a next_cursor line, json carries it as a key; a section too long for one page continues on the next, never dropped. Under profile, md Notes summarise each column - a long value list shows its 5 most frequent values - while json and yaml carry the statistics whole; for json/yaml a _corrupted field names any of the table's print files that failed to parse.

List a connection's tables, optionally narrowed by an fnmatch pattern over the dotted names. Use it to see which tables exist, their row counts, and whether each table's statistics are stale: detail: true adds each table's type, row_count, columns and profiled_at, and a freshness verdict - live, stale, or dormant when profiled_at is unreadable - with its age_days and the max_age_days it is judged against. A table whose threshold cannot be resolved carries threshold_error instead of a verdict. To find columns rather than tables, use search_columns; for the print's raw table index, get_manifest. Paged: every reply carries the total matched, and next_cursor while more remain.

Find columns across every table - the first call when the table holding a fact is not yet known. Filters, all optional and ANDed: a name glob, a text search over column names and their notes, and classification/sql_type/sensitivity/looks_like/redacted globs and a candidate_key match - for example every column holding an email address (looks_like: email) or contact details (sensitivity: contact). Read a table found this way with get_table_context; list tables rather than columns with list_tables. A match on a table read in part carries its scope block with rows_scanned/row_count, so a caller can tell a scanned-set number or key from a table-wide one; a match carrying a looks_like verdict carries the sampled/matched draw behind it, since a verdict from two values reads identically to one from ten thousand otherwise. A match carries sensitivity/redacted/candidate_key (and candidate_key_exception where the ratio falls short of 1.0) whenever the column does, so filtering on any of them returns the matched category, not just a bare column name. Paged: every reply carries the total matched, and next_cursor while more remain. A result carries unreadable_tables only when a table's own statistics or annotations failed to parse, on its first page: a statistics failure drops that table's columns from matches entirely; an annotations-only failure still returns them, without the annotation. A looks_like search also carries switched_off_tables on its first page, naming each table whose looks_like detection was switched off, where a non-match is not a finding.

Resolve a phrase, a code or a spelling against one column's published values, before writing a literal into a filter whose stored spelling get_table_context does not already list in full. Answers, in order of preference: stored - the text is a listed value, or folds to one, and the reply carries the spelling a predicate must use; definition - the text names what a value's note says it means; nearest - the listed values closest to the text, ranked; none; or unavailable where the column publishes no value list, lost it on the latest print, or redacts it. Every reply carries the column's coverage and how many values the print listed, plus a caveat sentence wherever that list is a sample - a spelling absent from a sample is not evidence it is absent from the column. An exhaustive list of at most fifty values rides along whole as domain, so a small vocabulary needs one call; on a table read in part, exhaustive is false, the reply carries the table's scope and row_count, and domain is the scanned rows' whole domain. Paged: every page repeats the scalar keys; the answer list comes before domain.

Return the print's index for a connection - its tables and the files each has, not a semantic catalogue of what they mean. Use it for an index field no other tool projects; for table names, row counts and staleness, list_tables with detail: true is shorter and already judges each table's freshness. The tables map is paged in FQN order with total and next_cursor; every other key of the document rides the first page whole, failed_tables filtered by pattern like tables.

Answer what changed between the database's last two prints: the recorded diff, whose summary counts every kind of change and whose changes list names each one - tables and columns added, removed or retyped, relationships changed, statistics that drifted. A statistic with no drift event did not change between the two runs; a table counted in unevaluated_tables was not compared. The changes list is paged in recorded order with total and next_cursor; every other key of the document rides the first page whole. A table's current state, rather than what changed, is get_table_context's answer.

Look up the dbprint format specification or the assertion DSL specification by section number, or the reading guide by heading: what a print's field means, or what a finding's spec_ref (e.g. '§2.2.4') refers to. A section returns its own text and lists its direct subsections, each read by its own number or heading; omit section for the heading tree. The guide is the one this dbprint version ships, not the copy stored with a print. Paged like get_table_context's md. Depends on no connection or print; what one print's tables hold is get_table_context's and list_tables' answer.