vs Kuzu

LoraDB vs Kuzu

LoraDB leads by 3.30× overall. Kuzu is an embedded columnar graph engine. Numbers from the same suite as the overview — identical fixtures, identical seed, identical iteration count.

81 comparable workloads1 omissions12 groups

3.30×slower

Kuzu vs LoraDB overall

Geometric-mean slowdown across the full suite.

54/ 81

Workloads won by LoraDB

Kuzu wins 18; 9 go to a third engine.

15close calls

Within ~10% on both sides

Workloads where neither engine ran away with the row.

Same suite as the overview — only the comparison engine differs.

Overall

LoraDB sets the row; Kuzu sits at 3.30× slower.

Geometric-mean slowdown across every workload both engines run. Empty rows happen when the language can’t express the workload — they’re called out below, not hidden.

LoraDBfastest

Kuzu3.30× slower

Per group

Wins by group.

For each group, the workload count, who wins it, and how the per-row tally breaks down between LoraDB, Kuzu, and any third engine that took a row.

setup1 workload · winner Grafeo
LoraDB2.70× slower
Kuzu700× slower
Rows · 10·0·1
writes8 workloads · winner Grafeo
LoraDB0.61× slower
Kuzu1.79× slower
Rows · 83·1·4
scans6 workloads · winner LoraDB
LoraDBfastest
Kuzu6.37× slower
Rows · 65·1
predicates12 workloads · winner LoraDB
LoraDBfastest
Kuzu1.14× slower
Rows · 128·2·2
strings5 workloads · winner Kuzu
LoraDB1.22× slower
Kuzufastest
Rows · 50·5
numerics6 workloads · winner Kuzu
LoraDB1.08× slower
Kuzufastest
Rows · 61·5
aggregates9 workloads · winner LoraDB
LoraDBfastest
Kuzu3.18× slower
Rows · 98·0·1
pipeline9 workloads · winner LoraDB
LoraDBfastest
Kuzu1.27× slower
Rows · 95·3·1
lists3 workloads · winner LoraDB
LoraDBfastest
Kuzu11.3× slower
Rows · 32·1
sort3 workloads · winner LoraDB
LoraDBfastest
Kuzu1.41× slower
Rows · 33·0
traversals15 workloads · winner LoraDB
LoraDBfastest
Kuzu20.4× slower
Rows · 1515·0
patterns4 workloads · winner LoraDB
LoraDBfastest
Kuzu2.80× slower
Rows · 44·0

Per workload

Raw timings, workload by workload.

Two columns — LoraDB and Kuzu — across every workload they share. Rows with no comparable Kuzurun are hidden here; they’re listed in the omissions block at the bottom.

setup(1)

WinnerGrafeo

Workload	Size	LoraDB	Kuzu	Winner
construct_empty	—	4.06 µs2.70× slower	1.05 ms700× slower	Grafeo

writes(8)

WinnerGrafeo

Workload	Size	LoraDB	Kuzu	Winner
bulk_edges	200	606.63 µsfastest	2.01 ms3.31× slower	LoraDB
bulk_set_match	1000	552.39 µs1.84× slower	300.03 µsfastest	Kuzu
delete_node	1000	355.70 µs1.47× slower	598.35 µs2.48× slower	Grafeo
merge_create	1000	112.57 µs1.25× slower	670.68 µs7.45× slower	Grafeo
merge_existing	1000	23.95 µs2.89× slower	162.06 µs19.6× slower	Grafeo
set_multiple_props	1000	21.15 µsfastest	169.72 µs8.03× slower	LoraDB
update_set	1000	18.01 µsfastest	188.62 µs10.5× slower	LoraDB
write_bulk	1000	1.46 ms1.93× slower	2.76 ms3.65× slower	Grafeo

scans(6)

WinnerLoraDB

Workload	Size	LoraDB	Kuzu	Winner
distinct	1000	198.38 µsfastest	448.06 µs2.26× slower	LoraDB
lookup_by_id	1000	716.64 nsfastest	116.07 µs162× slower	LoraDB
lookup_by_id_indexed	1000	684.00 nsfastest	116.18 µs170× slower	LoraDB
range_filter	1000	199.70 µs1.04× slower	192.68 µsfastest	Kuzu
scan_filtered	1000	149.08 µsfastest	158.49 µs1.06× slower	LoraDB
scan_label	1000	124.96 µsfastest	131.34 µs1.05× slower	LoraDB

predicates(12)

WinnerLoraDB

Workload	Size	LoraDB	Kuzu	Winner
where_compound_and_or	1000	213.07 µsfastest	230.83 µs1.08× slower	LoraDB
where_contains	1000	145.04 µsfastest	166.24 µs1.15× slower	LoraDB
where_ends_with	1000	143.60 µsfastest	167.41 µs1.17× slower	LoraDB
where_id_in_range	1000	142.71 µs2.02× slower	187.97 µs2.66× slower	Grafeo
where_in_list	1000	162.42 µsfastest	196.06 µs1.21× slower	LoraDB
where_modulo_eq	1000	127.32 µsfastest	167.83 µs1.32× slower	LoraDB
where_not	1000	166.05 µs1.01× slower	164.07 µsfastest	Kuzu
where_or	1000	147.16 µsfastest	185.19 µs1.26× slower	LoraDB
where_starts_with	1000	146.91 µsfastest	168.08 µs1.14× slower	LoraDB
where_string_gte	1000	180.97 µs1.07× slower	168.40 µsfastest	Kuzu
where_subexpr	1000	227.59 µs1.69× slower	194.85 µs1.45× slower	SurrealDB
where_two_props	1000	152.69 µsfastest	205.41 µs1.35× slower	LoraDB

strings(5)

WinnerKuzu

Workload	Size	LoraDB	Kuzu	Winner
string_concat	1000	186.55 µs1.27× slower	147.24 µsfastest	Kuzu
string_size	1000	172.17 µs1.10× slower	156.38 µsfastest	Kuzu
string_substring	1000	210.20 µs1.21× slower	173.57 µsfastest	Kuzu
string_to_lower	1000	201.00 µs1.33× slower	151.22 µsfastest	Kuzu
string_to_upper	1000	185.41 µs1.22× slower	151.96 µsfastest	Kuzu

numerics(6)

WinnerKuzu

Workload	Size	LoraDB	Kuzu	Winner
numeric_abs	1000	174.85 µs1.13× slower	154.99 µsfastest	Kuzu
numeric_ceil	1000	171.69 µs1.10× slower	155.40 µsfastest	Kuzu
numeric_floor	1000	175.48 µs1.10× slower	159.62 µsfastest	Kuzu
numeric_modulo	1000	144.95 µs1.02× slower	142.67 µsfastest	Kuzu
numeric_pow	1000	166.22 µs1.16× slower	143.34 µsfastest	Kuzu
numeric_round	1000	180.68 µsfastest	181.69 µs1.01× slower	LoraDB

aggregates(9)

WinnerLoraDB

Workload	Size	LoraDB	Kuzu	Winner
aggregate_avg	1000	81.43 µsfastest	253.54 µs3.11× slower	LoraDB
aggregate_collect	1000	78.90 µsfastest	275.13 µs3.49× slower	LoraDB
aggregate_count	1000	59.50 µs2.77× slower	250.00 µs11.7× slower	Grafeo
aggregate_count_distinct	1000	104.05 µsfastest	441.70 µs4.24× slower	LoraDB
aggregate_max	1000	79.97 µsfastest	263.33 µs3.29× slower	LoraDB
aggregate_min	1000	79.67 µsfastest	258.01 µs3.24× slower	LoraDB
aggregate_sum	1000	78.09 µsfastest	263.06 µs3.37× slower	LoraDB
grouped_aggregation	1000	154.46 µsfastest	549.19 µs3.56× slower	LoraDB
top_k	1000	186.50 µsfastest	248.58 µs1.33× slower	LoraDB

pipeline(9)

WinnerLoraDB

Workload	Size	LoraDB	Kuzu	Winner
case_when	1000	173.33 µsfastest	186.25 µs1.07× slower	LoraDB
coalesce_existing	1000	161.92 µsfastest	162.41 µs1.00× slower	LoraDB
computed_in_return	1000	151.15 µs1.07× slower	140.92 µsfastest	Kuzu
distinct_with_order	1000	508.01 µs2.08× slower	549.89 µs2.25× slower	Grafeo
predicate_via_function	1000	238.39 µs1.38× slower	172.38 µsfastest	Kuzu
with_aggregate_then_filter	1000	148.70 µsfastest	475.77 µs3.20× slower	LoraDB
with_distinct_then_count	1000	202.72 µsfastest	540.85 µs2.67× slower	LoraDB
with_pipeline	1000	186.41 µsfastest	333.60 µs1.79× slower	LoraDB
with_two_chained	1000	313.64 µs1.41× slower	222.76 µsfastest	Kuzu

lists(3)

WinnerLoraDB

Workload	Size	LoraDB	Kuzu	Winner
list_in_construction	1000	177.94 µs1.05× slower	168.89 µsfastest	Kuzu
list_unwind_explicit	1000	1.11 µsfastest	165.87 µs149× slower	LoraDB
range_function	1000	18.86 µsfastest	193.49 µs10.3× slower	LoraDB

sort(3)

WinnerLoraDB

Workload	Size	LoraDB	Kuzu	Winner
order_by_id_asc	1000	167.95 µsfastest	234.56 µs1.40× slower	LoraDB
order_by_multi_key	1000	211.99 µsfastest	276.69 µs1.31× slower	LoraDB
skip_limit	1000	162.19 µsfastest	251.28 µs1.55× slower	LoraDB

traversals(15)

WinnerLoraDB

Workload	Size	LoraDB	Kuzu	Winner
direct_record_traversal	500	831.70 nsfastest	255.86 µs308× slower	LoraDB
recursive_depth2	500	968.69 nsfastest	499.89 µs516× slower	LoraDB
recursive_depth3	500	1.04 µsfastest	540.33 µs520× slower	LoraDB
recursive_depth5	500	1.16 µsfastest	511.05 µs440× slower	LoraDB
relation_filter	500	117.78 µsfastest	406.36 µs3.45× slower	LoraDB
traversal_count_one_hop	500	59.36 µsfastest	288.00 µs4.85× slower	LoraDB
traversal_filter_one_hop	500	138.99 µsfastest	405.25 µs2.92× slower	LoraDB
traversal_one_hop	500	125.98 µsfastest	360.18 µs2.86× slower	LoraDB
traversal_reverse	500	123.67 µsfastest	365.12 µs2.95× slower	LoraDB
traversal_three_hop	500	236.10 µsfastest	1.13 ms4.80× slower	LoraDB
traversal_two_hop	500	165.58 µsfastest	649.69 µs3.92× slower	LoraDB
traversal_undirected	500	213.41 µsfastest	492.51 µs2.31× slower	LoraDB
variable_length_path	100	86.24 µsfastest	2.64 ms30.6× slower	LoraDB
varlen_2_to_5	100	123.74 µsfastest	3.90 ms31.5× slower	LoraDB
varlen_exact_5	100	56.89 µsfastest	3.86 ms67.9× slower	LoraDB

patterns(4)

WinnerLoraDB

Workload	Size	LoraDB	Kuzu	Winner
edge_subquery_clause	500	213.55 µsfastest	465.09 µs2.18× slower	LoraDB
star_fanout	1000	138.96 µsfastest	296.68 µs2.13× slower	LoraDB
star_fanout_count	1000	61.36 µsfastest	290.42 µs4.73× slower	LoraDB
star_fanout_filter	1000	112.53 µsfastest	312.37 µs2.78× slower	LoraDB

Honest omissions

Workloads Kuzu couldn’t run like-for-like.

Omissions

Workloads omitted only where the engine has no like-for-like equivalent — a missing scalar function, a different MERGE semantics, a reserved word. The reason is recorded with the workload, not silently dropped.

write_single
Kuzu
- Kuzu requires `CREATE NODE TABLE` before inserts; the empty fixture has no schema and adding one would change the iteration's measured cost.