Pulling specific metadata from files?

[JGB]

I've got a fairly good sized ebook library, and I've been cleaning it up lately, some things I prefer to set manually, but some fields I would like to have some info but am less fussy about it being perfect. The big one is published date, I don't really care if it's totally accurate, but it's nice for most books to have a general idea of when it was written so I can pick what I feel like reading.

Is there a way to have it pull just this data or specific columns of data from the file's meta-data without having it mess up all the series info I manually entered? I've tried on a few files and it overwrites the series.

[JSWolf]

The date of publish will be picked up if it's in the metadata. If not, you will have to enter it manually.

How are you updating your metadata such that is overwrites the series info?

[BetterRed]

Originally Posted by JGB

I've got a fairly good sized ebook library, and I've been cleaning it up lately, some things I prefer to set manually, but some fields I would like to have some info but am less fussy about it being perfect. The big one is published date, I don't really care if it's totally accurate, but it's nice for most books to have a general idea of when it was written so I can pick what I feel like reading.

Is there a way to have it pull just this data or specific columns of data from the file's meta-data without having it mess up all the series info I manually entered? I've tried on a few files and it overwrites the series.

@JGB - see Preferences ->Metadata download, here you can configure what columns you want to download. You can also configure what comes from each source via the Configure selected source button.

BR

[JGB]

Hi, thanks for the info,
however I am not trying to download the information from the internet, I'm trying to pull it from the books already existing metadata. The published column is not overwriting the series column, the series info from the metadata is overwriting my correct manually entered info with wrong info.
In the edit metadata mode I can load data from the book itself all or nothing.
Download mode works ok for selective data, but it takes forever vs pulling it from the metadata of the book itself when that's possible.

[BetterRed]

Aah - I don't think there is any way to selectively pull metadata elements from the format files, its all or nothing.

Are you interested in the pubdate of the edition you have, or the pubdate of the original, if the latter then I don't know of anyway to download or extract, I cut and paste from the viewer - if its there?

You could set the bulk metadata edit download off to just get pubdate from one source and let it run overnight. The metadata download speed is choked, this is to prevent calibre users overloading the source providers servers. If calibre didn't do that then they would complain to Kovid. They provide the information for free with no ad revenue, so they deserve a fair shake of the sauce bottle

BR

[JSWolf]

Originally Posted by JGB

Hi, thanks for the info,
however I am not trying to download the information from the internet, I'm trying to pull it from the books already existing metadata. The published column is not overwriting the series column, the series info from the metadata is overwriting my correct manually entered info with wrong info.
In the edit metadata mode I can load data from the book itself all or nothing.
Download mode works ok for selective data, but it takes forever vs pulling it from the metadata of the book itself when that's possible.

You are doing it backwards. Load the eBook and THEN fix the metadata. Don't have Calibre pull the metadata from the eBook once you've fixed things.

[JGB]

Originally Posted by JSWolf

You are doing it backwards. Load the eBook and THEN fix the metadata. Don't have Calibre pull the metadata from the eBook once you've fixed things.

That could work if there's a way to specify during the import to pull the data from the file names as the first priority, filling in data from the metadata present in the file only when that's missing, or to restrict the use of that metadata to only covers, publisher, published date and comments.
Because the metadata is not as accurate as the filename data for author and series, I can't permit caliber to pull only metadata from the internal information, it makes a mess of the imported books.
I don't need information from the online sources and that speed limit is annoying when the data is already accurate and present in the file names for most sections I care about(author, series) from the filenames, it's just a few pieces of missing information I'd like to pull from the file's metadata.

EDIT: Or is there a way to have Calibre use the file's metadata during import, and then over-write author and series from the filename data which I've already got corrected?

[JSWolf]

Originally Posted by JGB

That could work if there's a way to specify during the import to pull the data from the file names as the first priority, filling in data from the metadata present in the file only when that's missing, or to restrict the use of that metadata to only covers, publisher, published date and comments.
Because the metadata is not as accurate as the filename data for author and series, I can't permit caliber to pull only metadata from the internal information, it makes a mess of the imported books.
I don't need information from the online sources and that speed limit is annoying when the data is already accurate and present in the file names for most sections I care about(author, series) from the filenames, it's just a few pieces of missing information I'd like to pull from the file's metadata.

It doesn't matter where the data is coming from. You then fix the entry's information and use Polish Books to update the metadata and/or cover in the eBook. So you do get things fixed without having to worry about what the original metadata or filename is.

[JGB]

Originally Posted by JSWolf

It doesn't matter where the data is coming from. You then fix the entry's information and use Polish Books to update the metadata and/or cover in the eBook. So you do get things fixed without having to worry about what the original metadata or filename is.

That would seem to require a line by line editing of each individual book though, when there is already accurate data available, it's just split between two locations(filename and metadata).
Sorry if I'm being dense, I'm not very good at all of this, which is why I'm asking for help.

[BetterRed]

@JGB - You can extract the metadata from a format file into an OPF file with the ebook-meta command

With a bit of yourFavouriteTextEditor-fu you could wrangle/mangle the OPF into a CSV file and input that into the Import List plugin to update publisher and publication date.

The OPF file is XML, there are some command line XML2CSV converters around maybe you could make use of one of them.

Would be nice if the ebook-meta program had a --to-csv option

BR

[JSWolf]

Originally Posted by JGB

That would seem to require a line by line editing of each individual book though, when there is already accurate data available, it's just split between two locations(filename and metadata).
Sorry if I'm being dense, I'm not very good at all of this, which is why I'm asking for help.

Read the following and see if it helps.

http://manual.calibre-ebook.com/gui.html#add-books

[eschwartz]

Originally Posted by JGB

That could work if there's a way to specify during the import to pull the data from the file names as the first priority, filling in data from the metadata present in the file only when that's missing, or to restrict the use of that metadata to only covers, publisher, published date and comments.
Because the metadata is not as accurate as the filename data for author and series, I can't permit caliber to pull only metadata from the internal information, it makes a mess of the imported books.
I don't need information from the online sources and that speed limit is annoying when the data is already accurate and present in the file names for most sections I care about(author, series) from the filenames, it's just a few pieces of missing information I'd like to pull from the file's metadata.

EDIT: Or is there a way to have Calibre use the file's metadata during import, and then over-write author and series from the filename data which I've already got corrected?

I would like to be able to do this too. I want to set a custom column to equal the original filename.

However, at the current time it is impossible.

I am wondering now if I can do something with a custom calibre-debug script...