conditionnal custom column

[myki]

Greetings,

I am trying to build a custom column, which content must be different, depending on the series field, empty or not (for my kobo bookshelves).

Here is an example :
if {series} exists, i want : {author_sort} - {series}
if {series} is empty, i want : {author_sort}

Can you help me, please ?

Thanks in advance

[myki]

Hi Myki, here is the solution ! (yes i love to talk to myself, even if i'm not always agree with myself ).

{authors:'switch(count($, '&'), 1, $, list_item($, 0, '&'))'}{series:| - |}

If {series} is empty, the custom field will be feed with the first author only,
If {series} is not empty, the custom field will be feed with the first author, a "-", a space and the name of the serie.

ex : Rowling, JK - Harry Potter


(yes i love myself )

[eschwartz]

{author_sort}{series:| - |}

See: http://manual.calibre-ebook.com/temp...ced-formatting for the gory details and HOWTO.

EDIT: myki -- didn't see your post until I posted. But there really is no need to jump the gun and assume only the first author is wanted.

Also, you might want to consider reducing your own template to:

{authors:sublist(0,1,&)}{series:| - |}

Or for that matter,

{authors:list_item(0,&)}{series:| - |}

works fine too.

[myki]

your code is clear and perfect, thank you very much eschwartz
And to extract the firstname of the first author in author_sort, i use this :

{author_sort:re((.*?)\,(.*?)( &|),\2)}

Do you know a better code, more clear ??

[eschwartz]

Well, it always gets the first (and middle) name of the last author...

Try this:

{author_sort:re(^(.*?)\, (.*?)( & .*|$),\2)}

Anchoring the start and end of the string can often be a good idea also re() will keep going through the string to find as many matches as it can.

[myki]

Thanks for your help, I like it clear and efficient
I have (I hope) a last improvement to do :
I need to make a custom field (#kobocollections), which content is feed by another custom field if not empty (#categorieperso), and a custom value if empty ({authors:sublist(0,1,&)}{series:| - |}).

I succeeded by using :

program:
	a = field('#categorieperso');
	b = template('{authors:sublist(0,1,&)}{series:| - |}');
	test(field('#categorieperso'), a, b)

But can you explain to me why with this :

{#categorieperso:'ifempty($, template('{authors:sublist(0,1,&)}{series:| - |}'))'}

I get a template error formatter when my serie own a " ' " ??

Can you help me again, please ??

[chaley]

Originally Posted by myki

Thanks for your help, I like it clear and efficient
I have (I hope) a last improvement to do :
I need to make a custom field (#kobocollections), which content is feed by another custom field if not empty (#categorieperso), and a custom value if empty ({authors:sublist(0,1,&)}{series:| - |}).

I succeeded by using :

program:
	a = field('#categorieperso');
	b = template('{authors:sublist(0,1,&)}{series:| - |}');
	test(field('#categorieperso'), a, b)

FWIW: you could use

program:
	a = field('#categorieperso');
	b = template('{authors:sublist(0,1,&)}{series:| - |}');
	ifempty(a, b)

or

program:
	ifempty(
		field('#enum2'), 
		template('{authors:sublist(0,1,&)}{series:| - |}'))

But can you explain to me why with this :

{#categorieperso:'ifempty($, template('{authors:sublist(0,1,&)}{series:| - |}'))'}

I get a template error formatter when my serie own a " ' " ??

Can you help me again, please ??

You cannot put { or } characters in a template program mode template. The substitutions indicated by the {} characters happen before the template program fragment is evaluated, confusing the template evaluator. In your case the apostrophe is being interpreted as a string delimiter, creating a syntax error. You must use general program mode when you wish to evaluate sub-templates.

[myki]

I have adopted your first purpose, thanks again
Calibre is wonderful thanks to the author and to its community, thank you again for your help on this post and the others !

[eschwartz]

You don't strictly have to use GPM (when evaluating sub-templates), though, you *could* use TPM and substitute "[[" and "]]".
Granted, any such template is sufficiently complex enough to beg for GPM.

[chaley]

Originally Posted by eschwartz

You don't strictly have to use GPM (when evaluating sub-templates), though, you *could* use TPM and substitute "[[" and "]]".

Actually, you can't use [[ and ]] in myki's template. The problem is the "| - |" stuff inside the quotes.

<Technical>
The template parser runs in two passes. The first, mostly controlled by the python 'format' library, extracts and parses { ... } template information. The second, mostly controlled by my code, extracts TPM functions from inside the program string using the GPM parser. The problem that arises is that the first parser sees the ':| - |' after the #series and connects it with the nearest opening brace ({), which when using [[ and ]] is the one in front of #categorieperso. The result is a mess, causing the parser to throw up its hands.

If you leave the { and } around the series ({series:| - |}) then the parser gets it right, attaching the prefix and suffix to the right variable. Unfortunately, because this parser and evaluator run before the program parser, this sub-template is evaluated and replaced with the result. If that result contains a significant character such as a quote, comma, bracket, or bar then the resulting argument for the template function will be bogus.

I have considered removing the [[ and ]] hack because the rules for successful use approach incomprehensible.
</Technical>

[eschwartz]

Originally Posted by chaley

Actually, you can't use [[ and ]] in myki's template. The problem is the "| - |" stuff inside the quotes.

I knew that...

<Technical>
The template parser runs in two passes. The first, mostly controlled by the python 'format' library, extracts and parses { ... } template information. The second, mostly controlled by my code, extracts TPM functions from inside the program string using the GPM parser. The problem that arises is that the first parser sees the ':| - |' after the #series and connects it with the nearest opening brace ({), which when using [[ and ]] is the one in front of #categorieperso. The result is a mess, causing the parser to throw up its hands.

If you leave the { and } around the series ({series:| - |}) then the parser gets it right, attaching the prefix and suffix to the right variable. Unfortunately, because this parser and evaluator run before the program parser, this sub-template is evaluated and replaced with the result. If that result contains a significant character such as a quote, comma, bracket, or bar then the resulting argument for the template function will be bogus.

I have considered removing the [[ and ]] hack because the rules for successful use approach incomprehensible.
</Technical>

Yeah, if you cannot get it to work consistently, I agree it shouldn't be there. Just say template() and eval() only work in GPM.