|
| 1 | +[[pagination]] |
| 2 | += Pagination |
| 3 | +:page-aliases: pagination/index.adoc, pagination/offset-based.adoc, pagination/cursor-based.adoc, queries-aggregation/pagination/index.adoc, queries-aggregation/pagination/offset-based.adoc, queries-aggregation/pagination/cursor-based.adoc |
| 4 | + |
| 5 | + |
| 6 | +The Neo4j GraphQL Library offers two mechanisms for pagination: |
| 7 | + |
| 8 | +* Offset-based pagination - Pagination based on offsets, often associated with navigation via pages. |
| 9 | +* Cursor-based pagination - Pagination based on cursors, often associated with infinitely-scrolling applications. |
| 10 | + |
| 11 | + |
| 12 | +[[pagination-offset-based]] |
| 13 | +== Offset-based pagination |
| 14 | + |
| 15 | +Offset-based pagination, often associated with navigation via pages, can be achieved through the use of the `offset` and `limit` options available when querying for data. |
| 16 | + |
| 17 | +Using the following type definition: |
| 18 | + |
| 19 | +[source, graphql, indent=0] |
| 20 | +---- |
| 21 | +type User @node { |
| 22 | + name: String! |
| 23 | +} |
| 24 | +---- |
| 25 | + |
| 26 | +Fetch the first page of 10 by executing: |
| 27 | + |
| 28 | +[source, graphql, indent=0] |
| 29 | +---- |
| 30 | +query { |
| 31 | + users(limit: 10) { |
| 32 | + name |
| 33 | + } |
| 34 | +} |
| 35 | +---- |
| 36 | + |
| 37 | +And then on subsequent calls, introduce the `offset` argument and increment it by 10 on each call. |
| 38 | + |
| 39 | +Fetch the second page with: |
| 40 | + |
| 41 | +[source, graphql, indent=0] |
| 42 | +---- |
| 43 | +query { |
| 44 | + users(offset: 10, limit: 10) { |
| 45 | + name |
| 46 | + } |
| 47 | +} |
| 48 | +---- |
| 49 | + |
| 50 | +Fetch the third page with: |
| 51 | + |
| 52 | +[source, graphql, indent=0] |
| 53 | +---- |
| 54 | +query { |
| 55 | + users(offset: 20, limit: 10) { |
| 56 | + name |
| 57 | + } |
| 58 | +} |
| 59 | +---- |
| 60 | + |
| 61 | +And so on. |
| 62 | + |
| 63 | + |
| 64 | +=== Total number of pages |
| 65 | + |
| 66 | +You can fetch the total number of records for a certain type using its count query, and then divide that number by your entries per page in order to calculate the total number of pages. |
| 67 | +This determines what the last page is, and whether there is a next page. |
| 68 | + |
| 69 | +See xref::queries-aggregations/queries.adoc[Count] queries for details on how to execute these queries. |
| 70 | + |
| 71 | + |
| 72 | +=== Paginating relationship fields |
| 73 | + |
| 74 | +Say that in addition to the `User` type above, there is also a `Post` type which a `User` has many of. |
| 75 | +You can also fetch a `User` and then paginate through their posts: |
| 76 | + |
| 77 | +[source, graphql, indent=0] |
| 78 | +---- |
| 79 | +query { |
| 80 | + users(where: { |
| 81 | + name_EQ: "Billy" |
| 82 | + }) { |
| 83 | + name |
| 84 | + posts(offset: 20, limit: 10) { |
| 85 | + content |
| 86 | + } |
| 87 | + } |
| 88 | +} |
| 89 | +---- |
| 90 | + |
| 91 | + |
| 92 | +[[pagination-cursor-based]] |
| 93 | +== Cursor-based pagination |
| 94 | + |
| 95 | +On relationship fields, you are able to take advantage of cursor-based pagination, which is often associated with infinitely-scrolling applications. |
| 96 | + |
| 97 | +Using the following type definition: |
| 98 | + |
| 99 | +[source, graphql, indent=0] |
| 100 | +---- |
| 101 | +type User @node { |
| 102 | + name: String! |
| 103 | + posts: [Post!]! @relationship(type: "HAS_POST", direction: OUT) |
| 104 | +} |
| 105 | +
|
| 106 | +type Post @node { |
| 107 | + content: String! |
| 108 | +} |
| 109 | +---- |
| 110 | + |
| 111 | +If you wanted to fetch the posts of user "John Smith" 10 at a time, you would first fetch 10: |
| 112 | + |
| 113 | +[source, graphql, indent=0] |
| 114 | +---- |
| 115 | +query { |
| 116 | + users(where: { name_EQ: "John Smith" }) { |
| 117 | + name |
| 118 | + postsConnection(first: 10) { |
| 119 | + edges { |
| 120 | + node { |
| 121 | + content |
| 122 | + } |
| 123 | + } |
| 124 | + pageInfo { |
| 125 | + endCursor |
| 126 | + hasNextPage |
| 127 | + } |
| 128 | + } |
| 129 | + } |
| 130 | +} |
| 131 | +---- |
| 132 | + |
| 133 | +In the return value, if `hasNextPage` is `true`, you would pass `endCursor` into the next query of 10. |
| 134 | +You can do this with a variable, for example, `$after` in the following query: |
| 135 | + |
| 136 | +[source, graphql, indent=0] |
| 137 | +---- |
| 138 | +query Users($after: String) { |
| 139 | + users(where: { name_EQ: "John Smith" }) { |
| 140 | + name |
| 141 | + postsConnection(first: 10, after: $after) { |
| 142 | + edges { |
| 143 | + node { |
| 144 | + content |
| 145 | + } |
| 146 | + } |
| 147 | + pageInfo { |
| 148 | + endCursor |
| 149 | + hasNextPage |
| 150 | + } |
| 151 | + } |
| 152 | + } |
| 153 | +} |
| 154 | +---- |
| 155 | + |
| 156 | +You may continue until `hasNextPage` is `false` in the return - this is when you have reached the end of the data. |
| 157 | + |
| 158 | + |
| 159 | +=== `totalCount` |
| 160 | + |
| 161 | +The Connection fields also offer a `totalCount` field which can be used to calculate page numbers. |
| 162 | +This is useful if you want to paginate by cursor but use page numbers in your application. |
| 163 | + |
| 164 | +Add the `totalCount` field which returns the total number of results matching the filter used, for example: |
| 165 | + |
| 166 | +[source, graphql, indent=0] |
| 167 | +---- |
| 168 | +query Users($after: String) { |
| 169 | + users(where: { name_EQ: "John Smith" }) { |
| 170 | + name |
| 171 | + postsConnection(first: 10) { |
| 172 | + edges { |
| 173 | + node { |
| 174 | + content |
| 175 | + } |
| 176 | + } |
| 177 | + pageInfo { |
| 178 | + endCursor |
| 179 | + hasNextPage |
| 180 | + } |
| 181 | + totalCount |
| 182 | + } |
| 183 | + } |
| 184 | +} |
| 185 | +---- |
0 commit comments