mirror of
https://github.com/GreptimeTeam/greptimedb.git
synced 2026-05-28 10:50:39 +00:00
331 lines
35 KiB
HTML
331 lines
35 KiB
HTML
<!DOCTYPE html><html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width, initial-scale=1.0"><meta name="generator" content="rustdoc"><meta name="description" content="Source of the Rust file `src/common/function/src/aggrs/approximate/hll.rs`."><title>hll.rs - source</title><script>if(window.location.protocol!=="file:")document.head.insertAdjacentHTML("beforeend","SourceSerif4-Regular-6b053e98.ttf.woff2,FiraSans-Italic-81dc35de.woff2,FiraSans-Regular-0fe48ade.woff2,FiraSans-MediumItalic-ccf7e434.woff2,FiraSans-Medium-e1aa3f0a.woff2,SourceCodePro-Regular-8badfe75.ttf.woff2,SourceCodePro-Semibold-aa29a496.ttf.woff2".split(",").map(f=>`<link rel="preload" as="font" type="font/woff2"href="../../../../static.files/${f}">`).join(""))</script><link rel="stylesheet" href="../../../../static.files/normalize-9960930a.css"><link rel="stylesheet" href="../../../../static.files/rustdoc-17e0aaed.css"><meta name="rustdoc-vars" data-root-path="../../../../" data-static-root-path="../../../../static.files/" data-current-crate="common_function" data-themes="" data-resource-suffix="" data-rustdoc-version="1.96.0-nightly (ac7f9ec7d 2026-03-20)" data-channel="nightly" data-search-js="search-63369b7b.js" data-stringdex-js="stringdex-2da4960a.js" data-settings-js="settings-170eb4bf.js" ><script src="../../../../static.files/storage-41dd4d93.js"></script><script defer src="../../../../static.files/src-script-813739b1.js"></script><script defer src="../../../../src-files.js"></script><script defer src="../../../../static.files/main-5013f961.js"></script><noscript><link rel="stylesheet" href="../../../../static.files/noscript-f7c3ffd8.css"></noscript><link rel="alternate icon" type="image/png" href="../../../../static.files/favicon-32x32-eab170b8.png"><link rel="icon" type="image/svg+xml" href="../../../../static.files/favicon-044be391.svg"></head><body class="rustdoc src"><a class="skip-main-content" href="#main-content">Skip to main content</a><!--[if lte IE 11]><div class="warning">This old browser is unsupported and will most likely display funky things.</div><![endif]--><nav class="sidebar"><div class="src-sidebar-title"><h2>Files</h2></div></nav><div class="sidebar-resizer" title="Drag to resize sidebar"></div><main><section id="main-content" class="content" tabindex="-1"><div class="main-heading"><h1><div class="sub-heading">common_function/aggrs/approximate/</div>hll.rs</h1><rustdoc-toolbar></rustdoc-toolbar></div><div class="example-wrap digits-3"><pre class="rust"><code><a href=#1 id=1 data-nosnippet>1</a><span class="comment">// Copyright 2023 Greptime Team
|
|
<a href=#2 id=2 data-nosnippet>2</a>//
|
|
<a href=#3 id=3 data-nosnippet>3</a>// Licensed under the Apache License, Version 2.0 (the "License");
|
|
<a href=#4 id=4 data-nosnippet>4</a>// you may not use this file except in compliance with the License.
|
|
<a href=#5 id=5 data-nosnippet>5</a>// You may obtain a copy of the License at
|
|
<a href=#6 id=6 data-nosnippet>6</a>//
|
|
<a href=#7 id=7 data-nosnippet>7</a>// http://www.apache.org/licenses/LICENSE-2.0
|
|
<a href=#8 id=8 data-nosnippet>8</a>//
|
|
<a href=#9 id=9 data-nosnippet>9</a>// Unless required by applicable law or agreed to in writing, software
|
|
<a href=#10 id=10 data-nosnippet>10</a>// distributed under the License is distributed on an "AS IS" BASIS,
|
|
<a href=#11 id=11 data-nosnippet>11</a>// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
<a href=#12 id=12 data-nosnippet>12</a>// See the License for the specific language governing permissions and
|
|
<a href=#13 id=13 data-nosnippet>13</a>// limitations under the License.
|
|
<a href=#14 id=14 data-nosnippet>14</a>
|
|
<a href=#15 id=15 data-nosnippet>15</a></span><span class="doccomment">//! Two UDAFs are implemented for HyperLogLog:
|
|
<a href=#16 id=16 data-nosnippet>16</a>//!
|
|
<a href=#17 id=17 data-nosnippet>17</a>//! - `hll`: Accepts a string column and aggregates the values into a
|
|
<a href=#18 id=18 data-nosnippet>18</a>//! HyperLogLog state.
|
|
<a href=#19 id=19 data-nosnippet>19</a>//! - `hll_merge`: Accepts a binary column of states generated by `hll`
|
|
<a href=#20 id=20 data-nosnippet>20</a>//! and merges them into a single state.
|
|
<a href=#21 id=21 data-nosnippet>21</a>//!
|
|
<a href=#22 id=22 data-nosnippet>22</a>//! The states can be then used to estimate the cardinality of the
|
|
<a href=#23 id=23 data-nosnippet>23</a>//! values in the column by `hll_count` UDF.
|
|
<a href=#24 id=24 data-nosnippet>24</a>
|
|
<a href=#25 id=25 data-nosnippet>25</a></span><span class="kw">use </span>std::sync::Arc;
|
|
<a href=#26 id=26 data-nosnippet>26</a>
|
|
<a href=#27 id=27 data-nosnippet>27</a><span class="kw">use </span>common_query::prelude::<span class="kw-2">*</span>;
|
|
<a href=#28 id=28 data-nosnippet>28</a><span class="kw">use </span>common_telemetry::trace;
|
|
<a href=#29 id=29 data-nosnippet>29</a><span class="kw">use </span>datafusion::arrow::array::ArrayRef;
|
|
<a href=#30 id=30 data-nosnippet>30</a><span class="kw">use </span>datafusion::common::cast::{as_binary_array, as_string_array};
|
|
<a href=#31 id=31 data-nosnippet>31</a><span class="kw">use </span>datafusion::common::not_impl_err;
|
|
<a href=#32 id=32 data-nosnippet>32</a><span class="kw">use </span>datafusion::error::{DataFusionError, <span class="prelude-ty">Result </span><span class="kw">as </span>DfResult};
|
|
<a href=#33 id=33 data-nosnippet>33</a><span class="kw">use </span>datafusion::logical_expr::function::AccumulatorArgs;
|
|
<a href=#34 id=34 data-nosnippet>34</a><span class="kw">use </span>datafusion::logical_expr::{Accumulator <span class="kw">as </span>DfAccumulator, AggregateUDF};
|
|
<a href=#35 id=35 data-nosnippet>35</a><span class="kw">use </span>datafusion::prelude::create_udaf;
|
|
<a href=#36 id=36 data-nosnippet>36</a><span class="kw">use </span>datafusion_expr::Volatility;
|
|
<a href=#37 id=37 data-nosnippet>37</a><span class="kw">use </span>datatypes::arrow::datatypes::DataType;
|
|
<a href=#38 id=38 data-nosnippet>38</a><span class="kw">use </span>hyperloglogplus::{HyperLogLog, HyperLogLogPlus};
|
|
<a href=#39 id=39 data-nosnippet>39</a>
|
|
<a href=#40 id=40 data-nosnippet>40</a><span class="kw">use </span><span class="kw">crate</span>::utils::FixedRandomState;
|
|
<a href=#41 id=41 data-nosnippet>41</a>
|
|
<a href=#42 id=42 data-nosnippet>42</a><span class="kw">pub const </span>HLL_NAME: <span class="kw-2">&</span>str = <span class="string">"hll"</span>;
|
|
<a href=#43 id=43 data-nosnippet>43</a><span class="kw">pub const </span>HLL_MERGE_NAME: <span class="kw-2">&</span>str = <span class="string">"hll_merge"</span>;
|
|
<a href=#44 id=44 data-nosnippet>44</a>
|
|
<a href=#45 id=45 data-nosnippet>45</a><span class="kw">const </span>DEFAULT_PRECISION: u8 = <span class="number">14</span>;
|
|
<a href=#46 id=46 data-nosnippet>46</a>
|
|
<a href=#47 id=47 data-nosnippet>47</a><span class="kw">pub</span>(<span class="kw">crate</span>) <span class="kw">type </span>HllStateType = HyperLogLogPlus<String, FixedRandomState>;
|
|
<a href=#48 id=48 data-nosnippet>48</a>
|
|
<a href=#49 id=49 data-nosnippet>49</a><span class="kw">pub struct </span>HllState {
|
|
<a href=#50 id=50 data-nosnippet>50</a> hll: HllStateType,
|
|
<a href=#51 id=51 data-nosnippet>51</a>}
|
|
<a href=#52 id=52 data-nosnippet>52</a>
|
|
<a href=#53 id=53 data-nosnippet>53</a><span class="kw">impl </span>std::fmt::Debug <span class="kw">for </span>HllState {
|
|
<a href=#54 id=54 data-nosnippet>54</a> <span class="kw">fn </span>fmt(<span class="kw-2">&</span><span class="self">self</span>, f: <span class="kw-2">&mut </span>std::fmt::Formatter<<span class="lifetime">'_</span>>) -> std::fmt::Result {
|
|
<a href=#55 id=55 data-nosnippet>55</a> <span class="macro">write!</span>(f, <span class="string">"HllState<Opaque>"</span>)
|
|
<a href=#56 id=56 data-nosnippet>56</a> }
|
|
<a href=#57 id=57 data-nosnippet>57</a>}
|
|
<a href=#58 id=58 data-nosnippet>58</a>
|
|
<a href=#59 id=59 data-nosnippet>59</a><span class="kw">impl </span>Default <span class="kw">for </span>HllState {
|
|
<a href=#60 id=60 data-nosnippet>60</a> <span class="kw">fn </span>default() -> <span class="self">Self </span>{
|
|
<a href=#61 id=61 data-nosnippet>61</a> <span class="self">Self</span>::new()
|
|
<a href=#62 id=62 data-nosnippet>62</a> }
|
|
<a href=#63 id=63 data-nosnippet>63</a>}
|
|
<a href=#64 id=64 data-nosnippet>64</a>
|
|
<a href=#65 id=65 data-nosnippet>65</a><span class="kw">impl </span>HllState {
|
|
<a href=#66 id=66 data-nosnippet>66</a> <span class="kw">pub fn </span>new() -> <span class="self">Self </span>{
|
|
<a href=#67 id=67 data-nosnippet>67</a> <span class="self">Self </span>{
|
|
<a href=#68 id=68 data-nosnippet>68</a> <span class="comment">// Safety: the DEFAULT_PRECISION is fixed and valid
|
|
<a href=#69 id=69 data-nosnippet>69</a> </span>hll: HllStateType::new(DEFAULT_PRECISION, FixedRandomState::new()).unwrap(),
|
|
<a href=#70 id=70 data-nosnippet>70</a> }
|
|
<a href=#71 id=71 data-nosnippet>71</a> }
|
|
<a href=#72 id=72 data-nosnippet>72</a>
|
|
<a href=#73 id=73 data-nosnippet>73</a> <span class="doccomment">/// Create a UDF for the `hll` function.
|
|
<a href=#74 id=74 data-nosnippet>74</a> ///
|
|
<a href=#75 id=75 data-nosnippet>75</a> /// `hll` accepts a string column and aggregates the
|
|
<a href=#76 id=76 data-nosnippet>76</a> /// values into a HyperLogLog state.
|
|
<a href=#77 id=77 data-nosnippet>77</a> </span><span class="kw">pub fn </span>state_udf_impl() -> AggregateUDF {
|
|
<a href=#78 id=78 data-nosnippet>78</a> create_udaf(
|
|
<a href=#79 id=79 data-nosnippet>79</a> HLL_NAME,
|
|
<a href=#80 id=80 data-nosnippet>80</a> <span class="macro">vec!</span>[DataType::Utf8],
|
|
<a href=#81 id=81 data-nosnippet>81</a> Arc::new(DataType::Binary),
|
|
<a href=#82 id=82 data-nosnippet>82</a> Volatility::Immutable,
|
|
<a href=#83 id=83 data-nosnippet>83</a> Arc::new(<span class="self">Self</span>::create_accumulator),
|
|
<a href=#84 id=84 data-nosnippet>84</a> Arc::new(<span class="macro">vec!</span>[DataType::Binary]),
|
|
<a href=#85 id=85 data-nosnippet>85</a> )
|
|
<a href=#86 id=86 data-nosnippet>86</a> }
|
|
<a href=#87 id=87 data-nosnippet>87</a>
|
|
<a href=#88 id=88 data-nosnippet>88</a> <span class="doccomment">/// Create a UDF for the `hll_merge` function.
|
|
<a href=#89 id=89 data-nosnippet>89</a> ///
|
|
<a href=#90 id=90 data-nosnippet>90</a> /// `hll_merge` accepts a binary column of states generated by `hll`
|
|
<a href=#91 id=91 data-nosnippet>91</a> /// and merges them into a single state.
|
|
<a href=#92 id=92 data-nosnippet>92</a> </span><span class="kw">pub fn </span>merge_udf_impl() -> AggregateUDF {
|
|
<a href=#93 id=93 data-nosnippet>93</a> create_udaf(
|
|
<a href=#94 id=94 data-nosnippet>94</a> HLL_MERGE_NAME,
|
|
<a href=#95 id=95 data-nosnippet>95</a> <span class="macro">vec!</span>[DataType::Binary],
|
|
<a href=#96 id=96 data-nosnippet>96</a> Arc::new(DataType::Binary),
|
|
<a href=#97 id=97 data-nosnippet>97</a> Volatility::Immutable,
|
|
<a href=#98 id=98 data-nosnippet>98</a> Arc::new(<span class="self">Self</span>::create_merge_accumulator),
|
|
<a href=#99 id=99 data-nosnippet>99</a> Arc::new(<span class="macro">vec!</span>[DataType::Binary]),
|
|
<a href=#100 id=100 data-nosnippet>100</a> )
|
|
<a href=#101 id=101 data-nosnippet>101</a> }
|
|
<a href=#102 id=102 data-nosnippet>102</a>
|
|
<a href=#103 id=103 data-nosnippet>103</a> <span class="kw">fn </span>update(<span class="kw-2">&mut </span><span class="self">self</span>, value: <span class="kw-2">&</span>str) {
|
|
<a href=#104 id=104 data-nosnippet>104</a> <span class="self">self</span>.hll.insert(value);
|
|
<a href=#105 id=105 data-nosnippet>105</a> }
|
|
<a href=#106 id=106 data-nosnippet>106</a>
|
|
<a href=#107 id=107 data-nosnippet>107</a> <span class="kw">fn </span>merge(<span class="kw-2">&mut </span><span class="self">self</span>, raw: <span class="kw-2">&</span>[u8]) {
|
|
<a href=#108 id=108 data-nosnippet>108</a> <span class="kw">if let </span><span class="prelude-val">Ok</span>(serialized) = bincode::deserialize::<HllStateType>(raw)
|
|
<a href=#109 id=109 data-nosnippet>109</a> && <span class="kw">let </span><span class="prelude-val">Ok</span>(()) = <span class="self">self</span>.hll.merge(<span class="kw-2">&</span>serialized)
|
|
<a href=#110 id=110 data-nosnippet>110</a> {
|
|
<a href=#111 id=111 data-nosnippet>111</a> <span class="kw">return</span>;
|
|
<a href=#112 id=112 data-nosnippet>112</a> }
|
|
<a href=#113 id=113 data-nosnippet>113</a> <span class="macro">trace!</span>(<span class="string">"Warning: Failed to merge HyperLogLog from {:?}"</span>, raw);
|
|
<a href=#114 id=114 data-nosnippet>114</a> }
|
|
<a href=#115 id=115 data-nosnippet>115</a>
|
|
<a href=#116 id=116 data-nosnippet>116</a> <span class="kw">fn </span>create_accumulator(acc_args: AccumulatorArgs) -> DfResult<Box<<span class="kw">dyn </span>DfAccumulator>> {
|
|
<a href=#117 id=117 data-nosnippet>117</a> <span class="kw">let </span>data_type = acc_args.exprs[<span class="number">0</span>].data_type(acc_args.schema)<span class="question-mark">?</span>;
|
|
<a href=#118 id=118 data-nosnippet>118</a>
|
|
<a href=#119 id=119 data-nosnippet>119</a> <span class="kw">match </span>data_type {
|
|
<a href=#120 id=120 data-nosnippet>120</a> DataType::Utf8 => <span class="prelude-val">Ok</span>(Box::new(HllState::new())),
|
|
<a href=#121 id=121 data-nosnippet>121</a> other => <span class="macro">not_impl_err!</span>(<span class="string">"{HLL_NAME} does not support data type: {other}"</span>),
|
|
<a href=#122 id=122 data-nosnippet>122</a> }
|
|
<a href=#123 id=123 data-nosnippet>123</a> }
|
|
<a href=#124 id=124 data-nosnippet>124</a>
|
|
<a href=#125 id=125 data-nosnippet>125</a> <span class="kw">fn </span>create_merge_accumulator(acc_args: AccumulatorArgs) -> DfResult<Box<<span class="kw">dyn </span>DfAccumulator>> {
|
|
<a href=#126 id=126 data-nosnippet>126</a> <span class="kw">let </span>data_type = acc_args.exprs[<span class="number">0</span>].data_type(acc_args.schema)<span class="question-mark">?</span>;
|
|
<a href=#127 id=127 data-nosnippet>127</a>
|
|
<a href=#128 id=128 data-nosnippet>128</a> <span class="kw">match </span>data_type {
|
|
<a href=#129 id=129 data-nosnippet>129</a> DataType::Binary => <span class="prelude-val">Ok</span>(Box::new(HllState::new())),
|
|
<a href=#130 id=130 data-nosnippet>130</a> other => <span class="macro">not_impl_err!</span>(<span class="string">"{HLL_MERGE_NAME} does not support data type: {other}"</span>),
|
|
<a href=#131 id=131 data-nosnippet>131</a> }
|
|
<a href=#132 id=132 data-nosnippet>132</a> }
|
|
<a href=#133 id=133 data-nosnippet>133</a>}
|
|
<a href=#134 id=134 data-nosnippet>134</a>
|
|
<a href=#135 id=135 data-nosnippet>135</a><span class="kw">impl </span>DfAccumulator <span class="kw">for </span>HllState {
|
|
<a href=#136 id=136 data-nosnippet>136</a> <span class="kw">fn </span>update_batch(<span class="kw-2">&mut </span><span class="self">self</span>, values: <span class="kw-2">&</span>[ArrayRef]) -> DfResult<()> {
|
|
<a href=#137 id=137 data-nosnippet>137</a> <span class="kw">let </span>array = <span class="kw-2">&</span>values[<span class="number">0</span>];
|
|
<a href=#138 id=138 data-nosnippet>138</a>
|
|
<a href=#139 id=139 data-nosnippet>139</a> <span class="kw">match </span>array.data_type() {
|
|
<a href=#140 id=140 data-nosnippet>140</a> DataType::Utf8 => {
|
|
<a href=#141 id=141 data-nosnippet>141</a> <span class="kw">let </span>string_array = as_string_array(array)<span class="question-mark">?</span>;
|
|
<a href=#142 id=142 data-nosnippet>142</a> <span class="kw">for </span>value <span class="kw">in </span>string_array.iter().flatten() {
|
|
<a href=#143 id=143 data-nosnippet>143</a> <span class="self">self</span>.update(value);
|
|
<a href=#144 id=144 data-nosnippet>144</a> }
|
|
<a href=#145 id=145 data-nosnippet>145</a> }
|
|
<a href=#146 id=146 data-nosnippet>146</a> DataType::Binary => {
|
|
<a href=#147 id=147 data-nosnippet>147</a> <span class="kw">let </span>binary_array = as_binary_array(array)<span class="question-mark">?</span>;
|
|
<a href=#148 id=148 data-nosnippet>148</a> <span class="kw">for </span>v <span class="kw">in </span>binary_array.iter().flatten() {
|
|
<a href=#149 id=149 data-nosnippet>149</a> <span class="self">self</span>.merge(v);
|
|
<a href=#150 id=150 data-nosnippet>150</a> }
|
|
<a href=#151 id=151 data-nosnippet>151</a> }
|
|
<a href=#152 id=152 data-nosnippet>152</a> <span class="kw">_ </span>=> {
|
|
<a href=#153 id=153 data-nosnippet>153</a> <span class="kw">return </span><span class="macro">not_impl_err!</span>(
|
|
<a href=#154 id=154 data-nosnippet>154</a> <span class="string">"HLL functions do not support data type: {}"</span>,
|
|
<a href=#155 id=155 data-nosnippet>155</a> array.data_type()
|
|
<a href=#156 id=156 data-nosnippet>156</a> );
|
|
<a href=#157 id=157 data-nosnippet>157</a> }
|
|
<a href=#158 id=158 data-nosnippet>158</a> }
|
|
<a href=#159 id=159 data-nosnippet>159</a>
|
|
<a href=#160 id=160 data-nosnippet>160</a> <span class="prelude-val">Ok</span>(())
|
|
<a href=#161 id=161 data-nosnippet>161</a> }
|
|
<a href=#162 id=162 data-nosnippet>162</a>
|
|
<a href=#163 id=163 data-nosnippet>163</a> <span class="kw">fn </span>evaluate(<span class="kw-2">&mut </span><span class="self">self</span>) -> DfResult<ScalarValue> {
|
|
<a href=#164 id=164 data-nosnippet>164</a> <span class="prelude-val">Ok</span>(ScalarValue::Binary(<span class="prelude-val">Some</span>(
|
|
<a href=#165 id=165 data-nosnippet>165</a> bincode::serialize(<span class="kw-2">&</span><span class="self">self</span>.hll).map_err(|e| {
|
|
<a href=#166 id=166 data-nosnippet>166</a> DataFusionError::Internal(<span class="macro">format!</span>(<span class="string">"Failed to serialize HyperLogLog: {}"</span>, e))
|
|
<a href=#167 id=167 data-nosnippet>167</a> })<span class="question-mark">?</span>,
|
|
<a href=#168 id=168 data-nosnippet>168</a> )))
|
|
<a href=#169 id=169 data-nosnippet>169</a> }
|
|
<a href=#170 id=170 data-nosnippet>170</a>
|
|
<a href=#171 id=171 data-nosnippet>171</a> <span class="kw">fn </span>size(<span class="kw-2">&</span><span class="self">self</span>) -> usize {
|
|
<a href=#172 id=172 data-nosnippet>172</a> std::mem::size_of_val(<span class="kw-2">&</span><span class="self">self</span>.hll)
|
|
<a href=#173 id=173 data-nosnippet>173</a> }
|
|
<a href=#174 id=174 data-nosnippet>174</a>
|
|
<a href=#175 id=175 data-nosnippet>175</a> <span class="kw">fn </span>state(<span class="kw-2">&mut </span><span class="self">self</span>) -> DfResult<Vec<ScalarValue>> {
|
|
<a href=#176 id=176 data-nosnippet>176</a> <span class="prelude-val">Ok</span>(<span class="macro">vec!</span>[ScalarValue::Binary(<span class="prelude-val">Some</span>(
|
|
<a href=#177 id=177 data-nosnippet>177</a> bincode::serialize(<span class="kw-2">&</span><span class="self">self</span>.hll).map_err(|e| {
|
|
<a href=#178 id=178 data-nosnippet>178</a> DataFusionError::Internal(<span class="macro">format!</span>(<span class="string">"Failed to serialize HyperLogLog: {}"</span>, e))
|
|
<a href=#179 id=179 data-nosnippet>179</a> })<span class="question-mark">?</span>,
|
|
<a href=#180 id=180 data-nosnippet>180</a> ))])
|
|
<a href=#181 id=181 data-nosnippet>181</a> }
|
|
<a href=#182 id=182 data-nosnippet>182</a>
|
|
<a href=#183 id=183 data-nosnippet>183</a> <span class="kw">fn </span>merge_batch(<span class="kw-2">&mut </span><span class="self">self</span>, states: <span class="kw-2">&</span>[ArrayRef]) -> DfResult<()> {
|
|
<a href=#184 id=184 data-nosnippet>184</a> <span class="kw">let </span>array = <span class="kw-2">&</span>states[<span class="number">0</span>];
|
|
<a href=#185 id=185 data-nosnippet>185</a> <span class="kw">let </span>binary_array = as_binary_array(array)<span class="question-mark">?</span>;
|
|
<a href=#186 id=186 data-nosnippet>186</a> <span class="kw">for </span>v <span class="kw">in </span>binary_array.iter().flatten() {
|
|
<a href=#187 id=187 data-nosnippet>187</a> <span class="self">self</span>.merge(v);
|
|
<a href=#188 id=188 data-nosnippet>188</a> }
|
|
<a href=#189 id=189 data-nosnippet>189</a>
|
|
<a href=#190 id=190 data-nosnippet>190</a> <span class="prelude-val">Ok</span>(())
|
|
<a href=#191 id=191 data-nosnippet>191</a> }
|
|
<a href=#192 id=192 data-nosnippet>192</a>}
|
|
<a href=#193 id=193 data-nosnippet>193</a>
|
|
<a href=#194 id=194 data-nosnippet>194</a><span class="attr">#[cfg(test)]
|
|
<a href=#195 id=195 data-nosnippet>195</a></span><span class="kw">mod </span>tests {
|
|
<a href=#196 id=196 data-nosnippet>196</a> <span class="kw">use </span>datafusion::arrow::array::{BinaryArray, StringArray};
|
|
<a href=#197 id=197 data-nosnippet>197</a>
|
|
<a href=#198 id=198 data-nosnippet>198</a> <span class="kw">use </span>super::<span class="kw-2">*</span>;
|
|
<a href=#199 id=199 data-nosnippet>199</a>
|
|
<a href=#200 id=200 data-nosnippet>200</a> <span class="attr">#[test]
|
|
<a href=#201 id=201 data-nosnippet>201</a> </span><span class="kw">fn </span>test_hll_basic() {
|
|
<a href=#202 id=202 data-nosnippet>202</a> <span class="kw">let </span><span class="kw-2">mut </span>state = HllState::new();
|
|
<a href=#203 id=203 data-nosnippet>203</a> state.update(<span class="string">"1"</span>);
|
|
<a href=#204 id=204 data-nosnippet>204</a> state.update(<span class="string">"2"</span>);
|
|
<a href=#205 id=205 data-nosnippet>205</a> state.update(<span class="string">"3"</span>);
|
|
<a href=#206 id=206 data-nosnippet>206</a>
|
|
<a href=#207 id=207 data-nosnippet>207</a> <span class="kw">let </span>result = state.evaluate().unwrap();
|
|
<a href=#208 id=208 data-nosnippet>208</a> <span class="kw">if let </span>ScalarValue::Binary(<span class="prelude-val">Some</span>(bytes)) = result {
|
|
<a href=#209 id=209 data-nosnippet>209</a> <span class="kw">let </span><span class="kw-2">mut </span>hll: HllStateType = bincode::deserialize(<span class="kw-2">&</span>bytes).unwrap();
|
|
<a href=#210 id=210 data-nosnippet>210</a> <span class="macro">assert_eq!</span>(hll.count().trunc() <span class="kw">as </span>u32, <span class="number">3</span>);
|
|
<a href=#211 id=211 data-nosnippet>211</a> } <span class="kw">else </span>{
|
|
<a href=#212 id=212 data-nosnippet>212</a> <span class="macro">panic!</span>(<span class="string">"Expected binary scalar value"</span>);
|
|
<a href=#213 id=213 data-nosnippet>213</a> }
|
|
<a href=#214 id=214 data-nosnippet>214</a> }
|
|
<a href=#215 id=215 data-nosnippet>215</a>
|
|
<a href=#216 id=216 data-nosnippet>216</a> <span class="attr">#[test]
|
|
<a href=#217 id=217 data-nosnippet>217</a> </span><span class="kw">fn </span>test_hll_roundtrip() {
|
|
<a href=#218 id=218 data-nosnippet>218</a> <span class="kw">let </span><span class="kw-2">mut </span>state = HllState::new();
|
|
<a href=#219 id=219 data-nosnippet>219</a> state.update(<span class="string">"1"</span>);
|
|
<a href=#220 id=220 data-nosnippet>220</a> state.update(<span class="string">"2"</span>);
|
|
<a href=#221 id=221 data-nosnippet>221</a>
|
|
<a href=#222 id=222 data-nosnippet>222</a> <span class="comment">// Serialize
|
|
<a href=#223 id=223 data-nosnippet>223</a> </span><span class="kw">let </span>serialized = state.evaluate().unwrap();
|
|
<a href=#224 id=224 data-nosnippet>224</a>
|
|
<a href=#225 id=225 data-nosnippet>225</a> <span class="comment">// Create new state and merge the serialized data
|
|
<a href=#226 id=226 data-nosnippet>226</a> </span><span class="kw">let </span><span class="kw-2">mut </span>new_state = HllState::new();
|
|
<a href=#227 id=227 data-nosnippet>227</a> <span class="kw">if let </span>ScalarValue::Binary(<span class="prelude-val">Some</span>(bytes)) = <span class="kw-2">&</span>serialized {
|
|
<a href=#228 id=228 data-nosnippet>228</a> new_state.merge(bytes);
|
|
<a href=#229 id=229 data-nosnippet>229</a>
|
|
<a href=#230 id=230 data-nosnippet>230</a> <span class="comment">// Verify the merged state matches original
|
|
<a href=#231 id=231 data-nosnippet>231</a> </span><span class="kw">let </span>result = new_state.evaluate().unwrap();
|
|
<a href=#232 id=232 data-nosnippet>232</a> <span class="kw">if let </span>ScalarValue::Binary(<span class="prelude-val">Some</span>(new_bytes)) = result {
|
|
<a href=#233 id=233 data-nosnippet>233</a> <span class="kw">let </span><span class="kw-2">mut </span>original: HllStateType = bincode::deserialize(bytes).unwrap();
|
|
<a href=#234 id=234 data-nosnippet>234</a> <span class="kw">let </span><span class="kw-2">mut </span>merged: HllStateType = bincode::deserialize(<span class="kw-2">&</span>new_bytes).unwrap();
|
|
<a href=#235 id=235 data-nosnippet>235</a> <span class="macro">assert_eq!</span>(original.count(), merged.count());
|
|
<a href=#236 id=236 data-nosnippet>236</a> } <span class="kw">else </span>{
|
|
<a href=#237 id=237 data-nosnippet>237</a> <span class="macro">panic!</span>(<span class="string">"Expected binary scalar value"</span>);
|
|
<a href=#238 id=238 data-nosnippet>238</a> }
|
|
<a href=#239 id=239 data-nosnippet>239</a> } <span class="kw">else </span>{
|
|
<a href=#240 id=240 data-nosnippet>240</a> <span class="macro">panic!</span>(<span class="string">"Expected binary scalar value"</span>);
|
|
<a href=#241 id=241 data-nosnippet>241</a> }
|
|
<a href=#242 id=242 data-nosnippet>242</a> }
|
|
<a href=#243 id=243 data-nosnippet>243</a>
|
|
<a href=#244 id=244 data-nosnippet>244</a> <span class="attr">#[test]
|
|
<a href=#245 id=245 data-nosnippet>245</a> </span><span class="kw">fn </span>test_hll_batch_update() {
|
|
<a href=#246 id=246 data-nosnippet>246</a> <span class="kw">let </span><span class="kw-2">mut </span>state = HllState::new();
|
|
<a href=#247 id=247 data-nosnippet>247</a>
|
|
<a href=#248 id=248 data-nosnippet>248</a> <span class="comment">// Test string values
|
|
<a href=#249 id=249 data-nosnippet>249</a> </span><span class="kw">let </span>str_values = <span class="macro">vec!</span>[<span class="string">"a"</span>, <span class="string">"b"</span>, <span class="string">"c"</span>, <span class="string">"d"</span>, <span class="string">"e"</span>, <span class="string">"f"</span>, <span class="string">"g"</span>, <span class="string">"h"</span>, <span class="string">"i"</span>];
|
|
<a href=#250 id=250 data-nosnippet>250</a> <span class="kw">let </span>str_array = Arc::new(StringArray::from(str_values)) <span class="kw">as </span>ArrayRef;
|
|
<a href=#251 id=251 data-nosnippet>251</a> state.update_batch(<span class="kw-2">&</span>[str_array]).unwrap();
|
|
<a href=#252 id=252 data-nosnippet>252</a>
|
|
<a href=#253 id=253 data-nosnippet>253</a> <span class="kw">let </span>result = state.evaluate().unwrap();
|
|
<a href=#254 id=254 data-nosnippet>254</a> <span class="kw">if let </span>ScalarValue::Binary(<span class="prelude-val">Some</span>(bytes)) = result {
|
|
<a href=#255 id=255 data-nosnippet>255</a> <span class="kw">let </span><span class="kw-2">mut </span>hll: HllStateType = bincode::deserialize(<span class="kw-2">&</span>bytes).unwrap();
|
|
<a href=#256 id=256 data-nosnippet>256</a> <span class="macro">assert_eq!</span>(hll.count().trunc() <span class="kw">as </span>u32, <span class="number">9</span>);
|
|
<a href=#257 id=257 data-nosnippet>257</a> } <span class="kw">else </span>{
|
|
<a href=#258 id=258 data-nosnippet>258</a> <span class="macro">panic!</span>(<span class="string">"Expected binary scalar value"</span>);
|
|
<a href=#259 id=259 data-nosnippet>259</a> }
|
|
<a href=#260 id=260 data-nosnippet>260</a> }
|
|
<a href=#261 id=261 data-nosnippet>261</a>
|
|
<a href=#262 id=262 data-nosnippet>262</a> <span class="attr">#[test]
|
|
<a href=#263 id=263 data-nosnippet>263</a> </span><span class="kw">fn </span>test_hll_merge_batch() {
|
|
<a href=#264 id=264 data-nosnippet>264</a> <span class="kw">let </span><span class="kw-2">mut </span>state1 = HllState::new();
|
|
<a href=#265 id=265 data-nosnippet>265</a> state1.update(<span class="string">"1"</span>);
|
|
<a href=#266 id=266 data-nosnippet>266</a> <span class="kw">let </span>state1_binary = state1.evaluate().unwrap();
|
|
<a href=#267 id=267 data-nosnippet>267</a>
|
|
<a href=#268 id=268 data-nosnippet>268</a> <span class="kw">let </span><span class="kw-2">mut </span>state2 = HllState::new();
|
|
<a href=#269 id=269 data-nosnippet>269</a> state2.update(<span class="string">"2"</span>);
|
|
<a href=#270 id=270 data-nosnippet>270</a> <span class="kw">let </span>state2_binary = state2.evaluate().unwrap();
|
|
<a href=#271 id=271 data-nosnippet>271</a>
|
|
<a href=#272 id=272 data-nosnippet>272</a> <span class="kw">let </span><span class="kw-2">mut </span>merged_state = HllState::new();
|
|
<a href=#273 id=273 data-nosnippet>273</a> <span class="kw">if let </span>(ScalarValue::Binary(<span class="prelude-val">Some</span>(bytes1)), ScalarValue::Binary(<span class="prelude-val">Some</span>(bytes2))) =
|
|
<a href=#274 id=274 data-nosnippet>274</a> (<span class="kw-2">&</span>state1_binary, <span class="kw-2">&</span>state2_binary)
|
|
<a href=#275 id=275 data-nosnippet>275</a> {
|
|
<a href=#276 id=276 data-nosnippet>276</a> <span class="kw">let </span>binary_array = Arc::new(BinaryArray::from(<span class="macro">vec!</span>[
|
|
<a href=#277 id=277 data-nosnippet>277</a> bytes1.as_slice(),
|
|
<a href=#278 id=278 data-nosnippet>278</a> bytes2.as_slice(),
|
|
<a href=#279 id=279 data-nosnippet>279</a> ])) <span class="kw">as </span>ArrayRef;
|
|
<a href=#280 id=280 data-nosnippet>280</a> merged_state.merge_batch(<span class="kw-2">&</span>[binary_array]).unwrap();
|
|
<a href=#281 id=281 data-nosnippet>281</a>
|
|
<a href=#282 id=282 data-nosnippet>282</a> <span class="kw">let </span>result = merged_state.evaluate().unwrap();
|
|
<a href=#283 id=283 data-nosnippet>283</a> <span class="kw">if let </span>ScalarValue::Binary(<span class="prelude-val">Some</span>(bytes)) = result {
|
|
<a href=#284 id=284 data-nosnippet>284</a> <span class="kw">let </span><span class="kw-2">mut </span>hll: HllStateType = bincode::deserialize(<span class="kw-2">&</span>bytes).unwrap();
|
|
<a href=#285 id=285 data-nosnippet>285</a> <span class="macro">assert_eq!</span>(hll.count().trunc() <span class="kw">as </span>u32, <span class="number">2</span>);
|
|
<a href=#286 id=286 data-nosnippet>286</a> } <span class="kw">else </span>{
|
|
<a href=#287 id=287 data-nosnippet>287</a> <span class="macro">panic!</span>(<span class="string">"Expected binary scalar value"</span>);
|
|
<a href=#288 id=288 data-nosnippet>288</a> }
|
|
<a href=#289 id=289 data-nosnippet>289</a> } <span class="kw">else </span>{
|
|
<a href=#290 id=290 data-nosnippet>290</a> <span class="macro">panic!</span>(<span class="string">"Expected binary scalar values"</span>);
|
|
<a href=#291 id=291 data-nosnippet>291</a> }
|
|
<a href=#292 id=292 data-nosnippet>292</a> }
|
|
<a href=#293 id=293 data-nosnippet>293</a>
|
|
<a href=#294 id=294 data-nosnippet>294</a> <span class="attr">#[test]
|
|
<a href=#295 id=295 data-nosnippet>295</a> </span><span class="kw">fn </span>test_hll_merge_function() {
|
|
<a href=#296 id=296 data-nosnippet>296</a> <span class="comment">// Create two HLL states with different values
|
|
<a href=#297 id=297 data-nosnippet>297</a> </span><span class="kw">let </span><span class="kw-2">mut </span>state1 = HllState::new();
|
|
<a href=#298 id=298 data-nosnippet>298</a> state1.update(<span class="string">"1"</span>);
|
|
<a href=#299 id=299 data-nosnippet>299</a> state1.update(<span class="string">"2"</span>);
|
|
<a href=#300 id=300 data-nosnippet>300</a> <span class="kw">let </span>state1_binary = state1.evaluate().unwrap();
|
|
<a href=#301 id=301 data-nosnippet>301</a>
|
|
<a href=#302 id=302 data-nosnippet>302</a> <span class="kw">let </span><span class="kw-2">mut </span>state2 = HllState::new();
|
|
<a href=#303 id=303 data-nosnippet>303</a> state2.update(<span class="string">"2"</span>);
|
|
<a href=#304 id=304 data-nosnippet>304</a> state2.update(<span class="string">"3"</span>);
|
|
<a href=#305 id=305 data-nosnippet>305</a> <span class="kw">let </span>state2_binary = state2.evaluate().unwrap();
|
|
<a href=#306 id=306 data-nosnippet>306</a>
|
|
<a href=#307 id=307 data-nosnippet>307</a> <span class="comment">// Create a merge state and merge both states
|
|
<a href=#308 id=308 data-nosnippet>308</a> </span><span class="kw">let </span><span class="kw-2">mut </span>merge_state = HllState::new();
|
|
<a href=#309 id=309 data-nosnippet>309</a> <span class="kw">if let </span>(ScalarValue::Binary(<span class="prelude-val">Some</span>(bytes1)), ScalarValue::Binary(<span class="prelude-val">Some</span>(bytes2))) =
|
|
<a href=#310 id=310 data-nosnippet>310</a> (<span class="kw-2">&</span>state1_binary, <span class="kw-2">&</span>state2_binary)
|
|
<a href=#311 id=311 data-nosnippet>311</a> {
|
|
<a href=#312 id=312 data-nosnippet>312</a> <span class="kw">let </span>binary_array = Arc::new(BinaryArray::from(<span class="macro">vec!</span>[
|
|
<a href=#313 id=313 data-nosnippet>313</a> bytes1.as_slice(),
|
|
<a href=#314 id=314 data-nosnippet>314</a> bytes2.as_slice(),
|
|
<a href=#315 id=315 data-nosnippet>315</a> ])) <span class="kw">as </span>ArrayRef;
|
|
<a href=#316 id=316 data-nosnippet>316</a> merge_state.update_batch(<span class="kw-2">&</span>[binary_array]).unwrap();
|
|
<a href=#317 id=317 data-nosnippet>317</a>
|
|
<a href=#318 id=318 data-nosnippet>318</a> <span class="kw">let </span>result = merge_state.evaluate().unwrap();
|
|
<a href=#319 id=319 data-nosnippet>319</a> <span class="kw">if let </span>ScalarValue::Binary(<span class="prelude-val">Some</span>(bytes)) = result {
|
|
<a href=#320 id=320 data-nosnippet>320</a> <span class="kw">let </span><span class="kw-2">mut </span>hll: HllStateType = bincode::deserialize(<span class="kw-2">&</span>bytes).unwrap();
|
|
<a href=#321 id=321 data-nosnippet>321</a> <span class="comment">// Should have 3 unique values: "1", "2", "3"
|
|
<a href=#322 id=322 data-nosnippet>322</a> </span><span class="macro">assert_eq!</span>(hll.count().trunc() <span class="kw">as </span>u32, <span class="number">3</span>);
|
|
<a href=#323 id=323 data-nosnippet>323</a> } <span class="kw">else </span>{
|
|
<a href=#324 id=324 data-nosnippet>324</a> <span class="macro">panic!</span>(<span class="string">"Expected binary scalar value"</span>);
|
|
<a href=#325 id=325 data-nosnippet>325</a> }
|
|
<a href=#326 id=326 data-nosnippet>326</a> } <span class="kw">else </span>{
|
|
<a href=#327 id=327 data-nosnippet>327</a> <span class="macro">panic!</span>(<span class="string">"Expected binary scalar values"</span>);
|
|
<a href=#328 id=328 data-nosnippet>328</a> }
|
|
<a href=#329 id=329 data-nosnippet>329</a> }
|
|
<a href=#330 id=330 data-nosnippet>330</a>}
|
|
</code></pre></div></section></main></body></html> |