base_dictionary.md
In basetable: Fast and Memory-Efficient Base R Table Manipulation

basetable operation dictionary

| Family | Intent | Exported function | Base relationship | Backend relationship | Why not dplyr-style naming | | --- | --- | --- | --- | --- | --- | | Inspection | Compact structure preview | preview() | Extends str() with tabular focus | Formats metadata after standardization | preview() is descriptive; the tidyverse glimpse() name is avoided; summarise()-style names are unrelated | | Inspection | Dimensions | dims() | Thin wrapper around dim() | No special backend logic needed | Avoids turning shape inspection into a verb grammar | | Inspection | Column type summary | types() | Extends class() / typeof() | Reads normalized table metadata | Distinct from across()-style type tooling | | Inspection | First and last rows | headtail() | Combines head() and tail() | Uses ordinary row slicing | More compact than ad hoc preview verbs | | Inspection | Column summaries | describe() | Extends summary() | Computes summaries efficiently from normalized columns | Not a summarise() clone; oriented to EDA | | Rows | Row filtering | subset() | Direct base analogue | Evaluates rows then subsets through data.table | Keeps base semantics instead of filter() | | Columns | Keep columns | pick() | Mirrors positive column indexing | Validated column selection | select() is too strongly tied to dplyr | | Columns | Drop columns | drop() | Mirrors negative column indexing | Set difference over names | select(-x) is a dplyr idiom; drop() is explicit | | Columns | Rename columns | renamecols() | Extends names<-() with new = old pairs | data.table::setnames() | rename() is a dplyr verb name; renamecols() names the action explicitly | | Rows | Unique rows | uniquerows() | Direct extension of unique() | data.table's unique() | dplyr's distinct() name is avoided; pairs with duplicaterows() | | Transform | Add/modify columns | transform() | Direct base analogue | Mutable data.table assignment | Avoids mutate() grammar | | Transform | Block mutation/removal | within() | Direct base analogue | Evaluate then rebuild | Base name is already correct | | Order | Sort rows | reorder() | Extends order() to data frames | setorderv() handles multi-column ordering | Avoids SQL-flavored arrange() | | Grouping | Group summaries | aggregate() | Direct base analogue | Group aggregation in data.table | Avoids summarise() / grouped objects | | Grouping | Group counts | count() | Close to table() and row counts | .N in data.table | Kept as a narrow convenience, not a grammar anchor | | Join | Merge tables | merge() | Direct base analogue | merge.data.table() | Keeps established join naming | | Split/apply | Split data | split() | Direct base analogue | Group split in data.table | No grouped state object | | Split/apply | Apply by groups | by_apply() | Inspired by by() | Split then apply | Explicitly grouped without group_by() | | Split/apply | Recombine pieces | combine() | Natural complement to split | rbindlist() | Clear and base-like | | Reshape | Wide/long reshape | reshape() | Direct base analogue | Delegates to stats::reshape() | Base already has the right term | | Reshape | Stack columns | stack() | Direct base analogue | Delegates to utils::stack() | Base already has the right term | | Reshape | Unstack values | unstack() | Direct base analogue | Delegates to utils::unstack() | Base already has the right term | | Missingness | Missing summaries | missingness() | Extends is.na() / complete.cases() | Column/row scans after normalization | missing() would conflict with base R's primitive | | EDA | Frequency tables | freq() | Extends table() / prop.table() | Counts in data.table | Focused helper, not a grouped summarise layer | | EDA | Dataset profile | profile() | Extends summary() | Joins type, missingness, and distinct stats | EDA-oriented, not transformation-oriented | | EDA | Compare datasets | compare() | Extends all.equal()-style checking | Compares schema and summaries | Makes differences explicit without join grammar |