Search feature

[tuanngocnguyen]

CLI script

WIP: Still working on regex-match

Search text throughout the whole database.

Options:
--search=STRING                  String to search for.
--regex-match=STRING             Use regular expression to match the search string.
--tables=tablename:columnname    Tables and columns to search. Separate multiple tables/columns with a comma.
                                 If not specified, search all tables and columns.
                                 If specify table only, search all columns in the table.
                                 Example: 
                                    --tables=user:username,user:email
                                    --tables=user,assign_submission:submission
                                    --tables=user,assign_submission
--summary                        Summary mode, only shows column/table where the text is found.
                                 If not specified, run in detail mode, which shows the full text where the search string is found.
-h, --help                       Print out this help.

Example:
\$ sudo -u www-data /usr/bin/php admin/tool/advancedreplace/cli/find.php --search=thelostsoul --summary
\$ sudo -u www-data /usr/bin/php admin/tool/advancedreplace/cli/find.php --regex-match=thelostsoul --summary

[brendanheywood]

if you "" bash quote the params would this make the \\ just \ and read easier?

[tuanngocnguyen]

I will fix this

[brendanheywood]

can you please use either the php or the moodle apis for writing csv so this is escaped properly for content when it contains commas and quotes

[brendanheywood]

when we scale up I'm pretty sure this will start breaking as everything is in memory

So probably fine to merge right now, but we'll need to refactor this later. We will probably want to do things in two phases, the first phase returns a list of tables we will search, and then we'll want to do a search table by tables and that second layer will return a record set we can stream through

[brendanheywood]

also thinking ahead, we talked about this script just outputting to stdout, but it might be better if it outputs to a file and then we can output a set of progress bars to stdout. Knowing how long things will take and where we are up to with ETA's is going to be fairly important. Again not critical now but I'll make issues for this