• Can be an actual URL (e.g. http://www.msn.com)
• Can be a URL expression. This is expressed in JSON.
• Documented here

What can be done with URL Expressions:

• Merge results from multiple requests   More...

  Method: step_sequential (uses: stop_on, arguments, max_pass, max_pass_ok) Method: step_sequential_parse_json, step_sequential_parse_xml (uses: stop_on, arguments, max_pass, max_pass_ok, rows_path, columns, columns_path, column_names_path, join_set, join_sets) Method: enumerate_files (uses: folder_type, file_path, filename_regex_pattern, filename_columns)

  
  
• Set custom request headers   More...

  All methods support the "headers" array of strings (including "none" which can be used to pass headers only). Each header is of the form: "Header-Name: Header-Value". You cannot change/override the User-Agent header.

  
  
• Shape hierarchical data into tabular data   More...

  Method: parse_json, parse_xml (uses: rows_path, columns, columns_path, column_names_path, join_set, join_sets) Method: step_sequential_parse_json, step_sequential_parse_xml (uses: stop_on, arguments, max_pass, max_pass_ok, rows_path, columns, columns_path, column_names_path, join_set, join_sets)

  
  

• Expressions that identify single data values (e.g. numberfollowsnear, xpath, etc.)
• Can combine multiple with && separators
• Documented here

What can be done with queries:

• Narrow scope of document to examine   More...

  If multiple statements are provided (using the && separator), the results from the first are chained to the next and so on. As such, all functions technically support the narrowing of scope to locate a value.

  
  
• Target data with a characteristic   More...

  Includes most functions other than xpath and multilinexpath, as these could be considered more "address based". For example, numberfollowsnear looks for the characteristic of numeric text that's "after and closest" to a matching pattern.

  
  
• Target data with a specific "address"   More...

  Includes functions such as xpath and multilinexpath which look for data using a specific location in the source document, described using XPath.

  
  

• Expressions that identify table of choice (e.g. tablenumber, columnname, etc.)
• Can additionally transform tabular results (e.g. rowfilter, skiprowcount, etc.)
• Can combine multiple with && separators
• Documented here

What can be done with queries:

• Target a table with a characteristic   More...

  Includes functions: columnname, columnnameregex, isallowsblank, valueregex, and musthaverowfilter. These functions when used become requirements for a table in the source - the table cannot be returned if it fails these tests.

  
  
• Target a table with a specific "address"   More...

  Includes the "tablenumber" function only.

  
  
• Optionally transform the resulting table   More...

  Includes functions: rowfilter, allowedblankheadercolumns, skiprowcount, mergecolumndown.

  
  

• Can be a column name - implies FIRST() function
• Can be one of the following functions: FIRST, LAST, MAX, MIN, SUM, AVG (column name as a parameter)
• Can be a non-XS.QL expression whose syntax is documented here.

What can be done with column names:

• Select a single value from one or more rows in an already-identified table using an aggregation function   More...

  Functions FIRST, LAST, MAX, MIN, SUM, AVG attempted first, otherwise tries expression syntax.

  
  

• Strictly speaking this is not XS.QL but supports logical operators, comparisons, etc.
• Documented here

What can be done with value filters:

• Narrow the number of rows returned based on filter criteria

• Strictly speaking this is not XS.QL but supports multi-column sorting
• Documented here

What can be done with value sorting:

• Sort the tabular data on one or more columns (ascending or descending)