Feature request: index filtering

[leonardguy-dynes]

First of all, thanks heaps for such a useful tool. We've bought a license.

We have a collection of ephemeral indexes on our DB that are created and destroyed as required to tune query performance, based on an approach described at http://www.sqlnuggets.com/blog/sql-scripts-how-to-find-missing-indexes/

We'd like to exclude them from the generated IEntityTypeConfiguration files in the same way that we can exclude sprocs or tables, i.e. via something like:

FilterSettings.IndexFilters.Add(new RegexExcludeFilter("^IX_.+_[0-9A-F]{5}$"));

This would result in e.g. 30 fewer indexes appearing in the Configure() method for one of our main tables.

Thanks again for a great product.

[sjh37]

Thanks for buying a licence 💯

To get you going for now, look for the following function in the .ttinclude file. Add the two lines of code indicated by the arrows:

public List<RawIndex> ReadIndexes()
{
    if (DatabaseReaderPlugin != null)
        return DatabaseReaderPlugin.ReadIndexes();

    var result = new List<RawIndex>();
    using (var conn = _factory.CreateConnection())
    {
        if (conn == null)
            return result;

        conn.ConnectionString = Settings.ConnectionString;
        conn.Open();

        var cmd = GetCmd(conn);
        if (cmd == null)
            return result;

        var sql = IndexSQL();
        if (string.IsNullOrWhiteSpace(sql))
            return result;

        cmd.CommandText = sql + SpecialQueryFlags();

        using (var rdr = cmd.ExecuteReader())
        {
            var exclude = new Regex("^IX_.+_[0-9A-F]{5}$", RegexOptions.Compiled); // <<--------------------
            while (rdr.Read())
            {
                var index = new RawIndex
                (
                    rdr["TableSchema"].ToString().Trim(),
                    rdr["TableName"].ToString().Trim(),
                    rdr["IndexName"].ToString().Trim(),
                    ChangeType<byte>(rdr["KeyOrdinal"]),
                    rdr["ColumnName"].ToString().Trim(),
                    ChangeType<int>(rdr["ColumnCount"]),
                    ChangeType<bool>(rdr["IsUnique"]),
                    ChangeType<bool>(rdr["IsPrimaryKey"]),
                    ChangeType<bool>(rdr["IsUniqueConstraint"]),
                    ChangeType<int>(rdr["IsClustered"]) == 1
                );
                
                if(!exclude.IsMatch(index.IndexName)) // <<--------------------
                    result.Add(index);
            }
        }
    }

    return result;
}

A quick test shows this to be working ok 👍

[leonardguy-dynes]

Thanks, that approach is working for us