Lectures

We break down how to use JsonElement as a “JSON tree” in kotlinx.serialization: parse text via Json.parseToJsonElement(...), distinguish JsonObject/ JsonArray/ JsonPrimitive/ JsonNull, safely navigate nodes with when and as?, and also avoid confusing “the field is missing” with “the field exists, but is null”.

We break down the practice of safe navigation through a JSON tree in kotlinx.serialization: how to work with JsonElement/ JsonObject without crashes, why JsonElement? shows up everywhere, how “key is missing” differs from JsonNull, and how to package extraction into predictable helper functions ( getStringOrNull, getIntOrNull, getObjectOrNull, getArrayOrNull). Separately: safe conversions via toIntOrNull()/ toDoubleOrNull()/ toBooleanStrictOrNull() and common mistakes ( !!, toInt(), early defaults).

We will walk through a practical pipeline for processing incoming JSON in Kotlin: we separate parsing text into JsonElement from validating structure and business rules, design a result contract via ValidationResult ( Ok/ Error) instead of Boolean, write helper functions for safe field reading, and build an expense-import validator for a CLI (including fail-fast with the item index).

We break down how to manually assemble JSON reports in Kotlin via the JsonElement tree: compute aggregates separately, then pack them using buildJsonObject/ buildJsonArray, use JsonPrimitive and JsonNull, distinguish “no field” vs “field = null”, and also combine manual assembly with Json.encodeToJsonElement and get pretty output via prettyPrint.

We break down how to consciously choose between JSON and XML: “data vs document”, primitive typing ( true, 123, "text", null), lists, attributes, ordering, and schemas/validation. Using Kotlin examples, we show why a JsonElement tree is convenient for unstable structures, why it’s worth separating parsing from business validation, and why it’s better to generate via buildJsonObject rather than by concatenating strings.