What would be the best/correct way do query for

[Trel]

I have a custom read/unread column.
I'm aware of calibredb, but not very familiar with it.

What I tried to do was

calibredb --library-path=/path/to/library list --fields "series,series_index" --search 'series:True series_index:1.0 #readstatus:"=Read"' --sort-by 'series'

I got it mostly working, but three issues.

1. The sort is reverse alphabetical (I want alphabetical)
2. The sort is done where "insignificant" words like "A" or "The" are disregarded (I want them included)
3. It still prints the ID, despite me not including it in --fields (I don't want the ID shown).

I was able to handle this with a mixture of various Linux text processing tools like cut, sed, and sort, but ideally I wouldn't have to do this somehow?

Once I get that down, I want to additionally select the MAX series index for each series, but I have no clue how I'd do that.

[jmurphy]

It won't get rid of the id, but try:

list --fields series,series_index --search "series:true and series_index:1.0 and #readstatus:Read" --sort-by series --ascending

[Trel]

Originally Posted by jmurphy

It won't get rid of the id, but try:

list --fields series,series_index --search "series:true and series_index:1.0 and #readstatus:Read" --sort-by series --ascending

That addresses the reverse sort, but it also breaks filtering on readstatus because "read" matches "read" and "unread", it's gotta be "=Read". It unfortunately doesn't help the issue with 'a' and 'the' not getting sorted alphabetically either.

[kovidgoyal]

If you want to sort including leading articles create a custom column in the library that uses a template to generate the value of series. It will sort using the value as a simple string.

[jmurphy]

Originally Posted by kovidgoyal

If you want to sort including leading articles create a custom column in the library that uses a template to generate the value of series. It will sort using the value as a simple string.

Maybe a built-in Series Sort field similar to Author Sort and Title Sort would make sense and then allowing the user to sort by either the literal field or the sort field.

Is there any way to suppress the output of the id field?

[kovidgoyal]

Originally Posted by jmurphy

Maybe a built-in Series Sort field similar to Author Sort and Title Sort would make sense and then allowing the user to sort by either the literal field or the sort field.

Is there any way to suppress the output of the id field?

Too much UI overhead for a niche use case.

And no you cannot suppress ids without them you can get duplicate rows. If you want full control use --for-machine which will output json and then postp process however you like using a tool such as jq or similar.

[Trel]

I know it's considered poor form to hit the database directly, but I managed to get this to so far grab the latest book in all series, I'm going to incorporate my custom Read/Unread column soon. Is there any real harm/risk in opening the DB in read only mode and running selects?

SELECT
    series_id
    ,book_id
    ,MAX(series_index) AS series_index
    ,series
    ,title
FROM
(
    SELECT
        s.sort AS series
        ,b.id AS book_id
        ,b.title AS title
        ,s.id AS series_id
        ,b.series_index AS series_index
    FROM
        books_series_link AS bs
    LEFT JOIN books AS b
        ON bs.book = b.id
    LEFT JOIN series AS s
        ON bs.series = s.id
)
GROUP BY
    series_id
ORDER BY
    series ASC;

[chaley]

FWIW: a search for selecting the highest Series_Index ?

[Trel]

I'm trying to do this to be able to export things like CSV and similar.

My final query which looks to be working is

SELECT                                          
    series_id                                   
    ,book_id                                    
    ,MAX(series_index) AS series_index          
    ,series                                     
    ,title                                      
    ,read_status                                
FROM                                            
(                                               
    SELECT                                      
        s.sort AS series                        
        ,b.id AS book_id                        
        ,b.title AS title                       
        ,s.id AS series_id                      
        ,b.series_index AS series_index         
        ,cc2.value AS read_status               
    FROM                                        
        books_series_link AS bs                 
    LEFT JOIN books AS b                        
        ON bs.book = b.id                       
    LEFT JOIN series AS s                       
        ON bs.series = s.id                     
    LEFT join books_custom_column_2_link AS bcc2
        ON b.id = bcc2.book                     
    LEFT join custom_column_2 AS cc2            
        ON bcc2.value = cc2.id                  
)                                               
GROUP BY                                        
    series_id                                   
ORDER BY                                        
    series ASC;

I can add a WHERE clause to the outer query to filter on the read status

EDIT: this is all precursor for me searching out and finding if I've completed a completed series and adding custom metadata to that effect, but also something I frequently try to look up