In the Step seven – Separate user mentions, hashtags, and URLs section, we created new tables to hold the extracted hashtags, user mentions, and URLs, and then provided a way to link each row back to the original table via the id column. We followed the rules of database normalization by creating new tables that represent the one-to-many relationship between a tweet and user mentions, between a tweet and hashtags, or between a tweet and URLs. In this step, we will continue optimizing this table for performance and efficiency.

The column we are concerned with now is the query_phrase column. Looking at the column data, we can see that it contains the same phrases repeated over and over. These were apparently the search phrases that were originally used to locate and select the tweets that now exist in this dataset. Of the 498 tweets in the sentiment140 table, how many of the query phrases are repeated over and over? We can use the following SQL to detect...