Shape / columns · All operations

Change letter case

A variant of Clean up text

Convert text to lower, upper, title or sentence case.

Status

In arcform

Runs as a SQL step

arcform carries it as a sql: step today.

The DuckDB step

Lower, upper, title and sentence case

It reads the orders table the proofs share. Create it first:

SQL
CREATE TABLE orders AS SELECT * FROM (VALUES
  (1,  'Alice Ng',      'north', DATE '2026-01-03', 120.00, 'widget', 'a|b',   '{"channel":"web","items":[1,2]}'),
  (2,  'alice ng',      'north', DATE '2026-01-05',  80.50, 'gadget', 'b',     '{"channel":"shop","items":[3]}'),
  (3,  'Bob  Okafor',   'south', DATE '2026-01-05', 300.00, 'widget', 'a|c',   '{"channel":"web","items":[4,5,6]}'),
  (4,  'Carla Díaz',    'east',  DATE '2026-01-09',  NULL,  'gizmo',  NULL,    '{"channel":"web","items":[]}'),
  (5,  'Bob Okafor',    'south', DATE '2026-01-12',  45.25, 'gadget', 'c',     '{"channel":"phone","items":[7]}'),
  (6,  'Dev Patel',     'west',  DATE '2026-01-12', 510.00, 'widget', 'a',     '{"channel":"web","items":[8,9]}'),
  (7,  'Carla Diaz',    'east',  DATE '2026-01-20',  99.99, 'gizmo',  'b|c',   '{"channel":"shop","items":[10]}'),
  (8,  'Dev Patel',     'west',  DATE '2026-02-02', 510.00, 'widget', 'a',     '{"channel":"web","items":[8,9]}'),
  (9,  'Eve Larsen',    NULL,    DATE '2026-02-03',  15.00, 'gadget', NULL,    '{"channel":"web","items":[11]}'),
  (10, 'Alice Ng',      'north', DATE '2026-02-10', 220.00, 'gizmo',  'a|b|c', '{"channel":"phone","items":[12,13]}')
) t(order_id, customer, region, order_date, amount, product, tags, meta);

Then run the statement:

SQL
CREATE TEMP TABLE result AS
SELECT order_id, lower(customer) AS lowered, upper(customer) AS uppered,
       array_to_string(list_transform(string_split(lower(customer), ' '), lambda w: upper(w[1]) || w[2:]), ' ') AS titled,
       upper(customer[1]) || lower(customer[2:]) AS sentence
FROM orders;
SELECT * FROM result;
SELECT CASE WHEN ((SELECT [lowered, uppered, titled, sentence] FROM result WHERE order_id = 2) = ['alice ng', 'ALICE NG', 'Alice Ng', 'Alice ng'] AND (SELECT [lowered, uppered, titled, sentence] FROM result WHERE order_id = 3) = ['bob  okafor', 'BOB  OKAFOR', 'Bob  Okafor', 'Bob  okafor']) THEN 'as expected' ELSE error('alice ng and Bob  Okafor did not come out in lower, upper, title and sentence case') END AS verdict;
  • It needs nothing installed beyond DuckDB itself.
  • The statement checks its own result and raises an error when it is wrong: “alice ng and Bob Okafor did not come out in lower, upper, title and sentence case”.
  • It passed on DuckDB v2.0.0 (Cyanoptera) d54cbbe3df.

What other tools call it

Meridian's classification; an independent blind rating agreed on 95.7% of 324 rows (variant, version 3).

AWS Glue (1)
  • FormatCaseread 2026-09-27

    Change each string of a column to a given case

Duckle (1)
Ibis (3)
  • Capitalizeread 2026-09-27

    Upper-case the first letter of a string

  • Lowercaseread 2026-09-27

    Convert a string to lower case

  • Uppercaseread 2026-09-27

    Convert a string to upper case

KNIME Analytics Platform (1)
  • Case Converterread 2026-09-27

    Convert string columns to upper or lower case

OpenRefine (4)
  • Case conversionread 2026-09-27

    3 functions: toLowercase, toUppercase, toTitlecase

  • To lowercaseread 2026-09-27

    Convert the column's text to lower case

  • To titlecaseread 2026-09-27

    Convert the column's text to title case

  • To uppercaseread 2026-09-27

    Convert the column's text to upper case

RapidMiner Studio (1)
  • Transform Casesread 2026-09-27

    Convert all characters of a document to lower or upper case

Tableau Prep (3)
Trifacta (3)

Background

Case mapping converts a string to a particular form, uppercase, lowercase or titlecase, possibly for display to the user.[1] Case folding is mostly used for caseless comparison of text rather than to transform it, and Unicode bases it primarily on the lowercase mapping.[1] Case is often ignored in keyword searches, so that insignificant variations in capitalisation in queries and queried material do not matter.[2]

When to use it

Case mapping loses information: once the string Mark is lowercased, the original cannot be recovered, since it might have been mark or MARK.[1] Lowercase a copy for matching and keep the original spelling for display. (Meridian's view) Case-folded text should be used solely for internal processing, and generally should not be stored or displayed to the end user.[1]

Some case mappings depend on language: in Turkish, lowercase i maps to the dotted capital İ, and capital I maps to the dotless lowercase ı.[1] By default Unicode uppercases the German ß to SS, a mapping from one character to two.[1] In Python, lower leaves the word straße unchanged, while casefold turns it into strasse.[3]

Title case capitalises all words except minor ones such as articles and short prepositions, and its rules are standardised only at the level of house styles and style manuals.[2] Start case is a simplified variant of title case that, in text processing, usually capitalises all words irrespective of their part of speech.[2] Python's str.title treats any run of letters as a word, so apostrophes form word boundaries and they're becomes They'Re.[3] OpenRefine's toTitlecase takes a list of delimiter characters, so jean-michel jarre with a space and a hyphen as delimiters becomes Jean-Michel Jarre.[4] Title-casing names and addresses will get some of them wrong, so treat it as a display step and check the exceptions. (Meridian's view)

Sentence case capitalises the first word of a sentence, as well as proper nouns and other words a more specific rule requires.[2] Treat automatic sentence case as a rough display step, and check the proper nouns after it. (Meridian's view)

The method

Unicode defines three case-mapping properties for each character: uppercase, lowercase and titlecase.[2] Unicode's UnicodeData.txt holds the simple one-to-one case mappings, and SpecialCasing.txt adds context-sensitive, language-specific and one-to-many mappings, such as the one for uppercasing ß.[1] Full default case conversion includes context-sensitive rules, such as lowercasing the Greek capital sigma to final sigma at the end of a word.[1] Python's str.upper follows the Default Case Conversion algorithm of section 3.13.2 of the Unicode Standard.[3]

PostgreSQL's lower and upper convert a string by the rules of the database's locale.[5] PostgreSQL's initcap converts the first letter of each word to upper case and the rest to lower case, a word being a run of alphanumeric characters.[5] Unicode's titlecase mapping differs from its uppercase mapping chiefly for ligatures and digraphs, and the Dutch ij titlecases as IJ, so ijsje becomes IJsje.[1]

Compared with the alternatives

In DuckDB, 'hello' = 'hElLO' is false, while the same comparison with COLLATE NOCASE is true.[6]

Sources

  1. [1]FAQ - Character Properties, Case Mappings and Namesdocumentation
  2. [2]Letter case - Wikipediasecondary source
  3. [3]Built-in Types — Python 3.15.2 documentationdocumentation
  4. [4]GREL functions | OpenRefinedocumentation
  5. [5]PostgreSQL: Documentation: 18: 9.4. String Functions and Operatorsdocumentation
  6. [6]Collations – DuckDBdocumentation