df_08_marshalers

custom marshaler/unmarshaler example

this example demonstrates the Marshaler and Unmarshaler interfaces for types that need complete control over their binding and unbinding process.

key concepts demonstrated

custom marshaling/unmarshaling

• Marshaler interface: types define their own unbinding logic via MarshalDf()
• Unmarshaler interface: types define their own binding logic via UnmarshalDf()
• complete control: full control over the entire binding/unbinding process
• different from converters: converters handle type conversion, marshalers handle the entire process

use cases

• legacy data formats: handle data that doesn't map cleanly to struct fields
• complex validation: perform validation that requires access to the entire data map
• custom serialization: implement domain-specific serialization logic
• computed fields: fields that are calculated rather than stored directly

marshaler vs converter comparison

• converters: handle type conversion between specific types (string → Email)
• marshalers: handle the entire binding/unbinding process for a type
• converters: work on individual field values
• marshalers: work on complete data maps for the type

workflow demonstrated

1. custom time handling: demonstrate time parsing with multiple formats and validation
2. complex validation: show validation that requires multiple fields
3. legacy format support: handle data formats that don't match struct layout
4. computed fields: demonstrate fields calculated during marshaling/unmarshaling
5. round-trip compatibility: ensure data integrity through marshal/unmarshal cycles

example types

the example implements marshalers for:

• CustomTime: handles multiple time formats and timezone normalization
• ContactInfo: validates and normalizes contact information
• LegacyUser: handles legacy user data format conversion

usage

go run main.go

benefits

• complete control: full authority over binding/unbinding process
• validation integration: embed validation directly in the binding process
• legacy support: seamlessly handle legacy or non-standard data formats
• computed properties: dynamically calculate fields during binding/unbinding
• error context: provide detailed error information specific to the type