If you use conditional update, data in the table can be updated only when the conditions are met. If the conditions are not met, the update fails.

Prerequisites

  • The OTSClient instance is initialized. For more information, see Initialization.
  • A data table is created. Data is written to the table.

Usage notes

When you use the PutRow, UpdateRow, DeleteRow, or BatchWriteRow operation to update data, you can use conditional update to check row existence conditions and column-based conditions. The update is successful only when the conditions are met.

Conditional update includes row existence conditions and column-based conditions.

  • Row existence conditions: include IGNORE, EXPECT_EXIST, and EXPECT_NOT_EXIST.

    When you modify a table, the system first checks the row existence condition. If the row existence condition is not met, the modification fails, and an error is reported.

  • Column-based conditions: include SingleColumnCondition and CompositeCondition, which are used to determine whether the conditions are met based on the values of one or more columns.
    • SingleColumnCondition supports the comparison between a constant and a column that can be a primary key column. SingleColumnCondition does not support the comparison between two columns or two constants.
    • The inner node of CompositeCondition is used to perform logical operations. Subconditions can be SingleColumnCondition or CompositeCondition.

Conditional update can be used to implement optimistic locking. When you update a row, the value of the specified column is obtained. Assume that Column A has a value of 1. Obtain the value in Column A and set a condition that the value in Column A is 1. Update the value in Column A to 2. If the update fails, the row is updated by another client.

Limits

Column-based conditions for conditional update support relational operators such as =, !=, >, >=, <, and <=, and logical operators such as NOT, AND, and OR. You can specify a maximum of 10 conditions.

Parameters

Parameter Description
RowExistenceExpectation When you modify a table, the system first checks the row existence condition. If the row existence condition is not met, the modification fails, and an error is reported.

Row existence conditions include IGNORE, EXPECT_EXIST, and EXPECT_NOT_EXIST. In Tablestore, RowExistenceExpectation_IGNORE indicates IGNORE, RowExistenceExpectation_EXPECT_EXIST indicates EXPECT_EXIST, and RowExistenceExpectation_EXPECT_NOT_EXIST indicates EXPECT_NOT_EXIST.

  • IGNORE: No existence check is performed.
  • EXPECT_EXIST: The row is expected to exist. If the row exists, the condition is met. If the row does not exist, the condition is not met.
  • EXPECT_NOT_EXIST: Expect that the row does not exist. If the row does not exist, the condition is met. If the row exists, the condition is not met.
columnName The column name.
columnValue The comparison value of the column.
comparator The relational operator that is used to compare column values. For more information, see ComparatorType.

Relational operators include EQUAL (=), NOT_EQUAL (!=), GREATER_THAN (>), GREATER_EQUAL (>=), LESS_THAN (<), and LESS_EQUAL (<=). In Tablestore, CT_EQUAL indicates EQUAL (=), CT_NOT_EQUAL indicates NOT_EQUAL (!=), CT_GREATER_THAN indicates GREATER_THAN (>), CT_GREATER_EQUAL indicates GREATER_EQUAL (>=), CT_LESS_THAN indicates LESS_THAN (<), and CT_LESS_EQUAL indicates LESS_EQUAL (<=).

combinator The logical operator that is used to combine multiple conditions. For more information, see LogicalOperator.

Logical operators include NOT, AND, and OR. In Tablestore, LO_NOT indicates NOT, LO_AND indicates AND, and LO_OR indicates OR.

Different logical operators require different number of subconditions.
  • When the logical operator is NOT, only one subcondition can be added.
  • When the logical operator is AND or OR, you must add at least two subconditions.
filterIfMissing Specifies whether to pass the conditional check when a column does not exist in a row. The type of this parameter value is Boolean. The default value is true, which indicates that if the column does not exist in a row, the conditional check is passed, and the row meets the update conditions.

When filterIfMissing is set to false, if the column does not exist in a row, the conditional check fails, and the row does not meet the update conditions.

latestVersionOnly Specifies whether to use only the value of the latest version when the comparison value has multiple versions. The type of this parameter value is Boolean. The default value is true, which indicates that if the column has multiple versions of data, only the value of the latest version is used for comparison.

When latestVersionOnly is set to false, the comparison values of all versions are used for comparison if the comparison value has multiple versions. In this case, if only the value of a version meets the condition, the conditional check is passed, and the row meets the update conditions.

Examples

Update a row based on the specified primary key. If the specified row exists, the value of the name column is 'john', and the value of the addr column is 'china', the update succeeds. Otherwise, the update fails.

   var params = {
    tableName: "sampleTable",
    primaryKey: [{ 'gid': Long.fromNumber(20013) }, { 'uid': Long.fromNumber(20013) }],
    updateOfAttributeColumns: [{ 'PUT': [{ 'col1': 'test6' }] }]
};

// Specify that the expected row exists, that the value of the name column is 'john', and that the value of the addr column is 'china'.
var condition = new TableStore.CompositeCondition(TableStore.LogicalOperator.AND);
condition.addSubCondition(new TableStore.SingleColumnCondition('name', 'john', TableStore.ComparatorType.EQUAL));
condition.addSubCondition(new TableStore.SingleColumnCondition('addr', 'china', TableStore.ComparatorType.EQUAL));

params.condition = new TableStore.Condition(TableStore.RowExistenceExpectation.EXPECT_EXIST, condition);

client.updateRow(params,
    function (err, data) {
        if (err) {
            console.log('error:', err);
            return;
        }
        console.log('success:', data);
    });