Last change
on this file since 33868 was 33823, checked in by ak19, 4 years ago |
Recommitting mongo-data folder with renamed files with numbering.
|
File size:
2.4 KB
|
Line | |
---|
1 | /*
|
---|
2 |
|
---|
3 | The websites that have some MRI detected AND which are either in NZ or with NZ TLD
|
---|
4 | or (so if they're from overseas) don't contain /mi or mi.* in URL path.
|
---|
5 | We'll include Australia, to get the valid "kiwiproperty.com" website,
|
---|
6 | otherwise the sole exception, included in the result list.
|
---|
7 |
|
---|
8 |
|
---|
9 | db.getCollection('Websites').find({$and: [
|
---|
10 | {numPagesContainingMRI: {$gt: 0}},
|
---|
11 | {$or: [{geoLocationCountryCode: /(NZ|AU)/}, {domain: /\.nz$/}, {urlContainsLangCodeInPath: false}]}
|
---|
12 | ]}).count()
|
---|
13 |
|
---|
14 | 397
|
---|
15 |
|
---|
16 | Aggregate results by a count of country codes:
|
---|
17 |
|
---|
18 | db.Websites.aggregate([
|
---|
19 | {
|
---|
20 | $match: {
|
---|
21 | $and: [
|
---|
22 | {numPagesContainingMRI: {$gt: 0}},
|
---|
23 | {$or: [{geoLocationCountryCode: /(NZ|AU)/}, {domain: /\.nz$/}, {urlContainsLangCodeInPath: false}]}
|
---|
24 | ]
|
---|
25 | }
|
---|
26 | },
|
---|
27 | { $unwind: "$geoLocationCountryCode" },
|
---|
28 | {
|
---|
29 | $group: {
|
---|
30 | _id: {$toLower: '$geoLocationCountryCode'},
|
---|
31 | count: { $sum: 1 }
|
---|
32 | }
|
---|
33 | },
|
---|
34 | { $sort : { count : -1} }
|
---|
35 | ]);
|
---|
36 |
|
---|
37 | */
|
---|
38 |
|
---|
39 | /* 1 */
|
---|
40 | {
|
---|
41 | "_id" : "us",
|
---|
42 | "count" : 181.0
|
---|
43 | }
|
---|
44 |
|
---|
45 | /* 2 */
|
---|
46 | {
|
---|
47 | "_id" : "nz",
|
---|
48 | "count" : 89.0
|
---|
49 | }
|
---|
50 |
|
---|
51 | /* 3 */
|
---|
52 | {
|
---|
53 | "_id" : "au",
|
---|
54 | "count" : 21.0
|
---|
55 | }
|
---|
56 |
|
---|
57 | /* 4 */
|
---|
58 | {
|
---|
59 | "_id" : "de",
|
---|
60 | "count" : 19.0
|
---|
61 | }
|
---|
62 |
|
---|
63 | /* 5 */
|
---|
64 | {
|
---|
65 | "_id" : "fr",
|
---|
66 | "count" : 17.0
|
---|
67 | }
|
---|
68 |
|
---|
69 | /* 6 */
|
---|
70 | {
|
---|
71 | "_id" : "nl",
|
---|
72 | "count" : 16.0
|
---|
73 | }
|
---|
74 |
|
---|
75 | /* 7 */
|
---|
76 | {
|
---|
77 | "_id" : "dk",
|
---|
78 | "count" : 8.0
|
---|
79 | }
|
---|
80 |
|
---|
81 | /* 8 */
|
---|
82 | {
|
---|
83 | "_id" : "ca",
|
---|
84 | "count" : 7.0
|
---|
85 | }
|
---|
86 |
|
---|
87 | /* 9 */
|
---|
88 | {
|
---|
89 | "_id" : "es",
|
---|
90 | "count" : 6.0
|
---|
91 | }
|
---|
92 |
|
---|
93 | /* 10 */
|
---|
94 | {
|
---|
95 | "_id" : "gb",
|
---|
96 | "count" : 5.0
|
---|
97 | }
|
---|
98 |
|
---|
99 | /* 11 */
|
---|
100 | {
|
---|
101 | "_id" : "cz",
|
---|
102 | "count" : 4.0
|
---|
103 | }
|
---|
104 |
|
---|
105 | /* 12 */
|
---|
106 | {
|
---|
107 | "_id" : "at",
|
---|
108 | "count" : 3.0
|
---|
109 | }
|
---|
110 |
|
---|
111 | /* 13 */
|
---|
112 | {
|
---|
113 | "_id" : "it",
|
---|
114 | "count" : 3.0
|
---|
115 | }
|
---|
116 |
|
---|
117 | /* 14 */
|
---|
118 | {
|
---|
119 | "_id" : "ro",
|
---|
120 | "count" : 3.0
|
---|
121 | }
|
---|
122 |
|
---|
123 | /* 15 */
|
---|
124 | {
|
---|
125 | "_id" : "il",
|
---|
126 | "count" : 2.0
|
---|
127 | }
|
---|
128 |
|
---|
129 | /* 16 */
|
---|
130 | {
|
---|
131 | "_id" : "ch",
|
---|
132 | "count" : 2.0
|
---|
133 | }
|
---|
134 |
|
---|
135 | /* 17 */
|
---|
136 | {
|
---|
137 | "_id" : "bg",
|
---|
138 | "count" : 1.0
|
---|
139 | }
|
---|
140 |
|
---|
141 | /* 18 */
|
---|
142 | {
|
---|
143 | "_id" : "sg",
|
---|
144 | "count" : 1.0
|
---|
145 | }
|
---|
146 |
|
---|
147 | /* 19 */
|
---|
148 | {
|
---|
149 | "_id" : "mx",
|
---|
150 | "count" : 1.0
|
---|
151 | }
|
---|
152 |
|
---|
153 | /* 20 */
|
---|
154 | {
|
---|
155 | "_id" : "ir",
|
---|
156 | "count" : 1.0
|
---|
157 | }
|
---|
158 |
|
---|
159 | /* 21 */
|
---|
160 | {
|
---|
161 | "_id" : "cn",
|
---|
162 | "count" : 1.0
|
---|
163 | }
|
---|
164 |
|
---|
165 | /* 22 */
|
---|
166 | {
|
---|
167 | "_id" : "ie",
|
---|
168 | "count" : 1.0
|
---|
169 | }
|
---|
170 |
|
---|
171 | /* 23 */
|
---|
172 | {
|
---|
173 | "_id" : "jp",
|
---|
174 | "count" : 1.0
|
---|
175 | }
|
---|
176 |
|
---|
177 | /* 24 */
|
---|
178 | {
|
---|
179 | "_id" : "fi",
|
---|
180 | "count" : 1.0
|
---|
181 | }
|
---|
182 |
|
---|
183 | /* 25 */
|
---|
184 | {
|
---|
185 | "_id" : "gr",
|
---|
186 | "count" : 1.0
|
---|
187 | }
|
---|
188 |
|
---|
189 | /* 26 */
|
---|
190 | {
|
---|
191 | "_id" : "ru",
|
---|
192 | "count" : 1.0
|
---|
193 | }
|
---|
194 |
|
---|
195 | /* 27 */
|
---|
196 | {
|
---|
197 | "_id" : "unknown",
|
---|
198 | "count" : 1.0
|
---|
199 | }
|
---|
200 |
|
---|
Note:
See
TracBrowser
for help on using the repository browser.