Add functions to efficiently serialize records

dissect/esedb/record.py

@@ -4,7 +4,7 @@
 import struct
 from binascii import hexlify
 from functools import lru_cache
-from typing import TYPE_CHECKING, Any, Optional
+from typing import TYPE_CHECKING, Any, Iterator, Optional
 
 from dissect.util.xmemoryview import xmemoryview
 
@@ -50,6 +50,9 @@
         column = self._table.column(attr)
         return self._data.get(column, raw)
 
+    def as_dict(self, raw: bool = False) -> dict[str, RecordValue]:
+        return self._data.as_dict(raw)
+
     def __getitem__(self, attr: str) -> RecordValue:
         return self.get(attr)
 
@@ -183,6 +186,31 @@
         if value is not None:
             return self._parse_value(column, value, tag_field)
 
+    def as_dict(self, raw: bool = False) -> dict[str, RecordValue]:
+        """Serialize the record as a dictionary."""
+        obj = {}
+
+        def _iter_column_id() -> Iterator[Column]:
+            # Fixed
+            yield from range(1, self._last_fixed_id + 1)
+
+            # Variable
+            yield from range(128, self._last_variable_id + 1)
+
+            # Tagged
+            for idx in range(self._tagged_data_count):
+                yield self._get_tag_field(idx).identifier
+
+        for column_id in _iter_column_id():
+            column = self.table._column_id_map[column_id]
+
+            try:
+                obj[column.name] = self.get(column, raw)
+            except Exception as e:
+                obj[column.name] = f"!ERROR! {e}"
+
+        return obj
+
     def _parse_value(self, column: Column, value: bytes, tag_field: TagField = None) -> RecordValue:
         """Parse the raw value into the appropriate type.
 

tests/test_record.py

@@ -0,0 +1,40 @@
+from typing import BinaryIO
+
+from dissect.esedb.esedb import EseDB
+
+
+def test_as_dict(basic_db: BinaryIO):
+    db = EseDB(basic_db)
+    table = db.table("basic")
+
+    records = list(table.records())
+    assert len(records) == 2
+
+    assert [r.as_dict() for r in records] == [
+        {
+            "Id": 1,
+            "Bit": False,
+            "UnsignedByte": 213,
+            "Short": -1337,
+            "Long": -13371337,
+            "Currency": 1337133713371337,
+            "IEEESingle": 1.0,
+            "IEEEDouble": 13371337.13371337,
+            "DateTime": 4675210852477960192,
+            "UnsignedLong": 13371337,
+            "LongLong": -13371337,
+            "GUID": "3f360af1-6766-46dc-9af2-0dacf295c2a1",
+            "UnsignedShort": 1337,
+        },
+        {
+            "Id": 2,
+            "Bit": True,
+            "UnsignedByte": 255,
+            "Short": 1339,
+            "Long": 13391339,
+            "Currency": -1339133913391339,
+            "IEEESingle": -2.0,
+            "IEEEDouble": -13391339.13391339,
+            "DateTime": -4537072128574357504,
+        },
+    ]