@@ -24,7 +24,7 @@ or eccentric enumeration cases.
24
24
25
25
The marshalling penalty can be eliminated by setting ``coerce `` to
26
26
``False ``. This will require the developer to manually coerce the
27
- `` EnumField ` ` value to an Enum _ type object and is therefore usually
27
+ :class: ` ~django_enum.fields. EnumField ` value to an Enum _ type object and is therefore usually
28
28
not recommended - but may be appropriate if the dominate use case involves
29
29
high volume serialization to a raw value instead.
30
30
@@ -40,7 +40,7 @@ There is an obvious storage improvement when using a single column bit mask inst
40
40
we achieve better query performance as well? The following benchmarks compare storage and query
41
41
performance between boolean columns and bit masks.
42
42
43
- **Using a flag ** `` EnumField ` ` **out performs boolean columns in both
43
+ **Using a flag ** :class: ` ~django_enum.fields. EnumField ` **out performs boolean columns in both
44
44
storage and query performance in most scenarios. **
45
45
46
46
.. note ::
@@ -57,7 +57,7 @@ storage and query performance in most scenarios.**
57
57
No Indexing
58
58
-----------
59
59
60
- When no indexes are used, the flag `` EnumField ` ` saves a significant amount
60
+ When no indexes are used, the flag :class: ` ~django_enum.fields. EnumField ` saves a significant amount
61
61
of space over the boolean column model. The following plot shows the storage
62
62
efficiency improvement over boolean columns as the number of flags increases
63
63
for each supported RDBMS. The oracle line shows extents which are allocated in
@@ -94,13 +94,14 @@ does a full table scan:
94
94
Indexed Exact Queries
95
95
---------------------
96
96
97
- When an index is used, the flag `` EnumField ` ` marginally outperforms the
97
+ When an index is used, the flag :class: ` ~django_enum.fields. EnumField ` marginally outperforms the
98
98
boolean column equivalent in both storage and query performance for exact match
99
99
queries. It is also much simpler to define. When using the boolean column
100
100
approach a multi-column index must be used. By default PostgreSQL is compiled
101
101
with a maximum multi-column index size of 32 columns. This means that masks
102
102
with more than 32 flags must be split into multiple multi-column indexes which
103
- will further degrade performance compared to the equivalent flag ``EnumField ``.
103
+ will further degrade performance compared to the equivalent flag
104
+ :class: `~django_enum.fields.EnumField `.
104
105
105
106
The multi-column limit on MySQL is only 16 columns.
106
107
0 commit comments