datediff returns the number of days between two date columns. It's the go-to function for computing things like "days to ship", "days since signup", or "age in days". Spark 3.4 added a SQL-only alias date_diff that does the same thing — useful when you're writing a SQL expression but otherwise interchangeable.

date_add, date_sub, and add_months shift a date column forward or backward by a fixed number of days or months. They're the bread-and-butter functions for computing things like "30 days from order date", "one month before renewal", or "next billing cycle". Spark 3.4 added a one-word alias, dateadd, that's only reachable through expr().

date_part, datepart, and extract are generic ways to pull a single field out of a date, timestamp, or interval column. They cover the same ground as the dedicated functions like year, month, dayofmonth and hour, minute, second, but with one function call where the field name is a parameter — useful when the field you need is decided at runtime or driven by config.

Spark provides hour, minute, and second for pulling the time-of-day components out of a timestamp column. They're the time-side counterparts to year, month, and day and are useful for bucketing events by hour of day, filtering work hours, or building time-based features.

Spark provides a small family of functions for pulling individual date parts — year, month, day, week, quarter — out of a date or timestamp column. They're the building blocks for grouping by month, filtering by quarter, or partitioning a table by year.

Spark provides a small family of functions for getting the current date, time, and session timezone — useful for tagging records with a load time, calculating ages, or filtering on "today". They all return the value at the start of query evaluation, so every call within a single query sees the same value.

elt returns the n-th value from a list of column expressions, where n is a 1-based index. It's handy when you have an integer column that points at one of several sibling columns (or literal values) and you want to materialize the chosen value into a single result column.

bin returns the binary string representation of a long integer. conv is the more general tool — it converts a number string from one base to another, covering decimal, hex, octal, binary, or anything in between. Use them when you're working with bit flags, color codes, or any data where the representation matters as much as the value.

Beyond the familiar regexp_replace and regexp_extract, Spark 3.4 ships a set of regex helpers — regexp_count, regexp_extract_all, regexp_instr, regexp_like, and regexp_substr — that cover the common "count / extract-all / locate / test / grab-first" jobs that otherwise require chaining helpers or writing UDFs. They're SQL-only, so you reach them through expr() in Spark Scala.

url_encode converts a string into application/x-www-form-urlencoded format so it can be safely used in a URL. url_decode reverses that transformation. parse_url extracts pieces of a URL — the host, path, query string, or a specific query parameter. All three are Spark SQL functions, so you call them through expr().