[ 
https://issues.apache.org/jira/browse/MARMOTTA-632?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15251525#comment-15251525
 ] 

Sergio Fernández commented on MARMOTTA-632:
-------------------------------------------

Not actually necessary if the dataset is preprocessed (flattered) with a 
[script|https://gist.github.com/wikier/70043d9b2eed0b259b695ffdf9bc5d2b] like:

{code}
#!/bin/bash

if [ "$#" -ne 1 ]; then
    echo "missing directory"
    echo "Usage: ./flat_directory.sh DIR"
    exit -1
fi

DIR=$1

if [ ! -d "$DIR" ]; then
    echo "Directory '${DIR}' not found"
    exit -1
fi

TARGET="${DIR}-flat"
mkdir -p ${TARGET}
find ${DIR} -type f -print0 | while IFS= read -r -d $'\0' orig; do
    dest=${orig//\//_}
    dest=${dest/$DIR\_/$TARGET\/}
    echo "copying '${orig}' to '${dest}'..."
    cp ${orig} ${dest}
done
{code}

> KiWi Loader doesn't go through sub-directories
> ----------------------------------------------
>
>                 Key: MARMOTTA-632
>                 URL: https://issues.apache.org/jira/browse/MARMOTTA-632
>             Project: Marmotta
>          Issue Type: Bug
>          Components: KiWi Triple Store
>    Affects Versions: 3.3.0
>            Reporter: Sergio Fernández
>            Assignee: Sergio Fernández
>            Priority: Minor
>             Fix For: 3.4.0
>
>
> Some dataset dumps (e.g., {{DBpedia 2015-10}}) use a complex directories 
> structure, which currently is not very convenient to load with our bulk 
> loaders.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to