Appendix D — The verb you know

You arrive holding another tool’s word for it. This appendix is the index for that direction: find the name you know, read the god words beside it, and follow the link to the chapter that owns the idea. The other appendices run the opposite way, from god’s sentence outward.

The god column shows words, not whole pipelines, because most of these names map to a word plus a place to put it rather than to a verb alone.

D.1 From `dplyr`

you know	`god` says	the chapter
`filter()`	`keep`	Keeping rows
`select()`, `select(-x)`	`pick`, `all_but`	Choosing columns
`mutate()`	`add`	Adding a column
`arrange()`, `desc()`	`sort`, `descending`	Sorting and taking
`group_by()` + `summarise()`	`summarize ... by`	Summarizing by group
`slice_head(n)`, `slice_max()`	`take`, `sort` then `take`	Sorting and taking
`slice_tail(n)`	`take_last`, after a `sort`	Sorting and taking
`distinct()`	`pick` then `drop_duplicates`	Renaming and repeats
`rename(new = old)`	`rename`, same direction	Renaming and repeats
`pivot_longer()`	`lengthen`	Columns into rows
`pivot_wider()`	`widen`	Rows into columns
`left_join()`, `inner_join()`	`join`, `unmatched`	Another table
`semi_join()`, `anti_join()`	`keep(matching(...))`, with `not`	Has a partner
`bind_rows()`	`add_rows`	Adding rows
`complete()`, `expand()`	`add_combinations`, with `by`	The rows that are not there
`case_when()`	`when ... otherwise`	One way or another
`replace_values()`, `recode_values()`	`look_up`, with the ending written	One way or another
`coalesce()`	`first_present`	Missing values
`across(starts_with())`	`where`, `name`, `value`	One value per column
`row_number()`, `min_rank()`	`row_number`, `rank`	Each row a place
`cumsum()`, `lag()`, `lead()`	`running_total`, `previous`, `following`	The total so far
`sd()`	`standard_deviation`	Summarizing by group
`paste(collapse = )`, `stringr::str_flatten()`	`join_rows`, after a `sort`	Tidying text
`slider::slide_dbl()`, `zoo::rollmean()`	`rolling(average(...), n)`	The total so far
`tidyr::replace_na()`, `tidyr::drop_na()`	`fill_missing`, `drop_missing`	Missing values
`tidyr::fill()`	`latest`, after a `sort`	Missing values
`show_query()`	`show_as`, with seven targets	What it wrote
`collect()`	`collect`, the same word	Nothing runs until you ask

If you keep dplyr attached, four of the names in that column are also dplyr’s: collect, pick, rename and summarize. Three of them answer the same way whichever package you attached last, and pick is the one that does not. The names other packages take back is the whole of it.

D.2 From `pandas`

you know	`god` says	the chapter
`query()`, boolean mask	`keep`	Keeping rows
`[[...]]`, `drop(columns=)`	`pick`, `all_but`	Choosing columns
`assign()`	`add`	Adding a column
`sort_values(ascending=False)`	`sort`, `descending`	Sorting and taking
`groupby().agg()`	`summarize ... by`	Summarizing by group
`head(n)`, `nlargest()`	`take`, `sort` then `take`	Sorting and taking
`tail(n)`	`take_last`, after a `sort`	Sorting and taking
`rename(columns={"old": "new"})`	`rename`, the other direction	Renaming and repeats
`drop_duplicates()`	`drop_duplicates`, no subset	Renaming and repeats
`melt()`	`lengthen`	Columns into rows
`pivot_table()`, `pivot()`	`widen`	Rows into columns
`merge()`	`join`	Another table
`isin()` filter against a table	`matching`	Has a partner
`concat()`	`add_rows`	Adding rows
`reindex(MultiIndex.from_product())`	`add_combinations`	The rows that are not there
`np.select`, nested `where`	`when ... otherwise`	One way or another
`.replace({...})`, `.map({...})`	`look_up`, with the ending written	One way or another
`fillna()`, `dropna()`	`fill_missing`, `drop_missing`	Missing values
`ffill()`	`latest`, after a `sort`	Missing values
`cumsum()`, `shift()`	`running_total`, `previous`, `following`	The total so far
`std()`	`standard_deviation`	Summarizing by group
`.agg(", ".join)`, `.str.cat()`	`join_rows`, after a `sort`	Tidying text
`sort_values(na_position=)`	`sort ... missing first`, or the default last	Missing values
`rolling(7).mean()`	`rolling(average(...), 7)`	The total so far

D.3 From SQL

you know	`god` says	the chapter
`WHERE`	`keep`	Keeping rows
`SELECT a, b`	`pick`	Choosing columns
computed column in `SELECT`	`add`	Adding a column
`ORDER BY ... DESC`, `LIMIT`	`sort ... descending`, `take`	Sorting and taking
`GROUP BY` with aggregates	`summarize ... by`	Summarizing by group
`SELECT DISTINCT`	`pick` then `drop_duplicates`	Renaming and repeats
`JOIN ... ON`	`join ... by`	Another table
`EXISTS`, `IN (SELECT ...)`	`matching`	Has a partner
`UNION ALL`	`add_rows`	Adding rows
`CROSS JOIN` of two `DISTINCT`s, joined back	`add_combinations`	The rows that are not there
`CASE WHEN ... ELSE ... END`	`when ... otherwise`	One way or another
`COALESCE`	`first_present`	Missing values
`RANK() OVER`, `ROW_NUMBER()`	`rank`, `row_number`	Each row a place
`SUM() OVER (ORDER BY)`, `LAG`, `LEAD`	`running_total`, `previous`, `following`	The total so far
`STDDEV`	`standard_deviation`	Summarizing by group
`STRING_AGG`, `GROUP_CONCAT`, `LISTAGG`	`join_rows`, after a `sort`	Tidying text
`NULLS FIRST`, `NULLS LAST`	`sort ... missing first`, or the default last	Missing values
`LAST_VALUE(x IGNORE NULLS) OVER`	`latest`, after a `sort`	Missing values
`AVG() OVER (ROWS BETWEEN 6 PRECEDING AND CURRENT ROW)`	`rolling(average(...), 7)`	The total so far

One absence from all three tables is deliberate. There is no row for your tool’s way of applying an arbitrary function to a column, because god has no word for it. The scalar vocabulary is closed, and the honest translation of apply is the exit the last table in Appendix E points at.