Question: How do you store Japanese characters in DynamoDB?

Answer

Store Japanese characters directly in a DynamoDB String attribute. DynamoDB strings use UTF-8, so hiragana, katakana and kanji require no table character-set setting or special database configuration.

With a document client, an item such as { pk: "USER#123", displayName: "山田太郎", city: "東京" } stores those strings directly. With the low-level API, a text attribute is represented as {"S":"東京"}. Let the SDK handle request serialization instead of manually escaping every character or converting text to Binary.

Keep source files, input files and output tools configured for UTF-8. If retrieved text looks corrupted, inspect the original input and the terminal or application rendering before changing the database representation.

Size limits are measured in bytes, not displayed characters. Japanese text commonly takes more bytes per character than ASCII, so count the encoded size when approaching item or key limits. String sort order follows UTF-8 bytes rather than a Japanese linguistic collation.

If visually equivalent Unicode forms should match as keys, normalize consistently in the application before both writes and reads. See item sizing and sort-key behavior.

Source: AWS strings and naming rules.