[librecat-dev] Catmandu marc questions

[Johann Rolschewski]

Hi Lars,

> The example provided by /Johann,/ instead of doing the below, how one could give as input a txt file named OldNames.csv and the replacements be in a file named NewNames.csv?

you can use lookup tables as Patrick described in one of his last posts:

1. Create CSV files

names.csv:

oldname,newname
name1a,name1b
name2a,name2b
name3a,name3b

titles.csv:

oldtitle,newtitle
title1a,title1b
title2a,title2b
title3a,title3b

2. Create a fix file

replace.fix:

do marc_each()
    marc_map(100a,tmp.name);
    lookup(tmp.name,names.csv,delete:1)
    if exists(tmp.name)
        marc_set(100a,$.tmp.name)
        remove_field(tmp.name)
    end
    marc_map(245a,tmp.title);
    lookup(tmp.title,titles.csv,delete:1)
    if exists(tmp.title)
        marc_set(245a,$.tmp.title)
        remove_field(tmp.title)
    end
    remove_field(tmp)
end

3. Run conversion

$ catmandu convert MARC to MARC --fix replace.fix < input.mrc > output.mrc

> last, i do not know how to produce a clean report, 
> say i wanted to have a txt file with all 001 values, or a tag that is repeatable....
> 
> i noticed
> 
> catmandu convert --fix 'retain(_id)' MARC to YAML < file_0.mrc
> 
> has an output that adds for example ' character, when we have a digit
> 
> ---
> _id: '120090304134844'
> ...
> ---
> _id: recorda
> ...
> ---
> _id: recordb
> ...
> ---
> _id: recordc
> ...
> ---
> _id: recorde
> ...
> ---
> _id: '8797989999140813'
> 
> Is there a way to have just the values, in a row by row format?

I don't know any possibility to generate a CSV report while doing a MARC
to MARC conversion with Catmandu. You could use another fix and generate
CSV reports for the input and output and join them afterwards.

4. Create a fix file:

report.fix:

marc_map(100a,name)
marc_map(245a,title)

5. Run conversions:

$ catmandu convert MARC to CSV --fix report.fix --fields _id,name,title
< input.mrc > input.csv

$ catmandu convert MARC to CSV --fix report.fix --fields _id,name,title
< output.mrc > output.csv

6. Generate report:

$ join -t , input.csv output.csv > report.csv


Johann