MySQL CHAR_LENGTH vs. LENGTH: Bytes, Unicode and Spaces

In MySQL, CHAR_LENGTH() counts code points in a nonbinary string, while LENGTH() counts its bytes. Compare both on the same expression. A larger byte count can be normal for Unicode text; neither result is a general count of visible characters. These examples use the MySQL 8.4 rules documented in the string-function reference.

Compare the exact Unicode input

The following are expected results for nonbinary utf8mb4 strings, derived from the documentation and checked with Unicode code-point and UTF-8 byte arithmetic. They are not results from a MySQL server test.

The utf8mb4 character set supports supplementary characters such as this emoji. The two accented forms can look alike while containing different code points. Unicode grapheme boundaries describe user-perceived characters; a combining mark can belong to the same visible character as its base. Do not normalize input merely to make a count agree: that changes the text being measured.

Run a scoped comparison

With a UTF-8 encoded query file or client input, this read-only statement gives an expected code-point count of 1 and byte count of 2:

SELECT CHAR_LENGTH(_utf8mb4'é') AS code_points,
       LENGTH(_utf8mb4'é') AS bytes;

The character-set introducer labels the following literal's encoding. It does not convert incorrectly encoded bytes into UTF-8. Check how your client sends the query before using an accented literal as a diagnostic.

Check the column and retrieved spaces

In MySQL, VARCHAR(n) declares a character limit. Its effective capacity also depends on the row's byte limits and character set. A byte count alone therefore does not prove a value will fit, and expression bytes are not a measurement of total disk storage.

For values within the declared limit, VARCHAR retains trailing spaces. CHAR normally removes them on retrieval unless PAD_CHAR_TO_FULL_LENGTH is enabled. A length function measures the expression it receives, after that retrieval behavior. Excess trailing spaces beyond a VARCHAR limit can instead be truncated with a warning. The CHAR and VARCHAR documentation explains these exceptions. Inspect the original input and retrieved value before trimming anything.

Use the String Length guide to choose a counting unit, or the Unicode normalization guide to inspect composed and decomposed text. When working in another database, check its own rules; the PostgreSQL varchar guide covers a separate engine.