Create and use a secondary index-Tablestore(Tablestore)-阿里云帮助中心

Prerequisites

The Tablestore CLI installed and configured with access credentials. For more information, see Tablestore CLI.
A data table with predefined columns. The maximum number of versions (max versions) must be 1. For more information, see Manage data tables.

Create a global or local secondary index for a data table.

create_index -t <tableName> -n <indexName> -i <indexType> --pk <primaryKeyName> --attr <definedColumn>

Parameter	Required	Description
-t, --table	No	The data table name. Not required if a data table is already selected with `use`.
-n, --name	Yes	The secondary index name.
-i, --index_type	No	The secondary index type. Valid values: global (default): global secondary index. The indexed columns and primary key columns from the data table are asynchronously replicated to the index table. Synchronization latency is typically in milliseconds. local: local secondary index. Data is synchronously written to the index table. Data is available for queries immediately after it is written to the data table.
-k, --pk	Yes	The indexed columns of the index table, consisting of primary key columns and predefined columns from the data table. Separate multiple columns with commas. For a local secondary index, the first primary key column of the index table must be the same as the first primary key column of the data table.
-a, --attr	Yes	The attribute columns of the index table, which are predefined columns from the data table. Separate multiple columns with commas.
-b, --without_base_data	No	Excludes existing data from the data table when the secondary index is created. By default, existing data is included.

Examples

Create a global secondary index named index0 for the data table mytable, including existing data.
```
create_index -t mytable -n index0 -i global --pk uid,pid -a name,col0
```
Create a global secondary index named index1 for the data table mytable, excluding existing data.
```
create_index -t mytable -n index1 -i global --pk uid,pid -a name,col0 -b
```
Create a local secondary index named index2 for the data table mytable.
```
create_index -t mytable -n index2 -i local --pk uid,name -a col0,col1
```

Select an index table for subsequent query and export operations.

use --wc -t <indexName>

Parameter	Required	Description
--wc	No	Indicates that the specified table is a data table or index table rather than a search index.
-t, --table	Yes	The index table name.

Examples

Use the index table index0.

use -t index0

Run the desc command to view information about a data table and its secondary indexes.

desc -t <tableName>

Parameter	Required	Description
-t, --table	No	The data table or index table name. Not required if a table is already selected with `use`.
-f, --print_format	No	The output format. Valid values: json (default) and table.
-o, --output	No	Outputs table information to a local JSON file.

Examples

View information about the current table.
```
desc
```
Save table information to a local file.
```
desc -o /tmp/describe_table_meta.json
```

Read a single row of data from an index table by primary key.

Note

If the specified row does not exist, an empty result is returned.

get --pk '<pk_values>'

Parameter	Required	Description
-k, --pk	Yes	The primary key values of the index table, expressed as a JSON array. The number and types of primary key columns must match those of the index table.
-c, --columns	No	The columns to read. Specify primary key columns or attribute columns. Separate multiple columns with commas. If not specified, the entire row is returned.
--max_version	No	The maximum number of versions to read.
--time_range_start	No	Reads data within a version number range. `time_range_start` and `time_range_end` specify the start and end timestamps respectively. The range is a left-closed, right-open interval: [start, end).
--time_range_end	No
--time_range_specific	No	Reads data of a specific version number.
-o, --output	No	Outputs query results to a local JSON file.

Examples

Read a row from the index table where the primary key values are "86" and 6771.

get --pk '["86", 6771]'

Scan an index table to retrieve all data or a specified number of rows.

scan --limit <n>

Parameter	Required	Description
--limit	No	The maximum number of rows to return. If not specified, all rows in the table are scanned.

Examples

Scan up to 10 rows from the index table.

scan --limit 10

Export data from an index table to a local JSON file.

scan -o /localpath/filename.json

Parameter	Required	Description
-c, --columns	No	The columns to export. Specify primary key columns or attribute columns. Separate multiple columns with commas. If not specified, the entire row is exported.
--max_version	No	The maximum number of versions to export.
--time_range_start	No	Exports data within a version number range. `time_range_start` and `time_range_end` specify the start and end timestamps respectively. The range is a left-closed, right-open interval: [start, end).
--time_range_end	No
--time_range_specific	No	Exports data of a specific version number.
--backward	No	Exports data in descending order of primary keys.
-o, --output	Yes	The local file path to export data to.
-l, --limit	No	The maximum number of rows to export.
-b, --begin	No	The start and end points of the export range. The primary key range is a left-closed, right-open interval: [start, end). Note If `--begin` is set to `[null,null]`, the start point is `[INF_MIN,INF_MIN]`, which means each primary key column starts from the minimum value. If `--end` is set to `[null,null]`, the end point is `[INF_MAX,INF_MAX]`, which means each primary key column ends at the maximum value.
-e, --end	No

Examples

Export all data from the index table to a local file.
```
scan -o /tmp/mydata.json
```
Export only the uid and name columns.
```
scan -o /tmp/mydata.json -c uid,name
```

Export data within a specified primary key range.

scan -o /tmp/mydata.json -b '["86", 6771]' -e '["86", 6775]'

Delete an index table that is no longer needed.

drop_index -t <tableName> -i <indexName> -y

Parameter	Required	Description
-t, --table	No	The data table name. Not required if a data table is already selected with `use`.
-i, --index	Yes	The secondary index name.
-y, --yes	Yes	Confirms the deletion without prompting.

Examples

Delete the index table index0 from the current data table.
```
drop_index -i index0 -y
```
Delete the index table index0 from the data table mytable.
```
drop_index -t mytable -i index0 -y
```