[question] how to set default value if a path value is empty

[aaronjfox]

Is there a straight forward way of returning a specified value if a path defined in the template is empty? I am using the package in an AWS Lambda function to parse XML docs to JSON, then upload that into DynamoDB. The issue I am having with camaro returning blank if a path is empty is DynamoDB doesn't allow an attribute value to contain an empty string. Currently I am looping through the object returned by transform to check for empty strings, but hoping there is a better way I may be missing. I am admittedly not well versed in XPath, but since 'if a path is empty, return blank' is listed as custom syntax I figured it best to ask here first.

Thanks in advance and thanks for a great package.

[tuananh]

you mean if the value is blank, return a default value?

[aaronjfox]

Yes. Either a default value for any empty element, or explicitly defining a default value for each element in the template if it does not return a value after parsing. Essentially:

xml = <someParent> <someChild>Hello World</someChild> <anotherChild>Hello Again</anotherChild> </someParent>

template = { parent: ['//someParent', { sChild: 'someChild', aChild: 'anotherChild', rChild: 'randomChild' }]}

output = { "parent": [ "sChild": "Hello World", "aChild": "Hello Again", "rChild": "" <------ Is there a way to keep this path from being an empty string if not found? ] }

Edit: Sorry for the formatting. Hopefully it makes sense, if not I can fix it later!

[tuananh]

there is a work around this using if else condition but it's very ugly

sth like this

const { transform } = require('camaro')

const xml = `
    <items>
        <item>hello</item>
        <item />
        <item>world</item>
    </items>
`

const template = {
    items: ['//items/item', `
        concat(
            text(),
            substring(
                "default val", 
                1, 
                number(not(text())) * string-length("default val")
            )
        )
    `]
}

;(async function main() {
    const output = await transform(xml, template)
    console.log(JSON.stringify(output, null, 4))
})();

output

{
    "items": [
        "hello",
        "default val",
        "world"
    ]
}

explanation: use text() value and then concat with the other default value if the condition is true.

so if text() has value => the condition is false => text() + empty => text() still

if text() is empty => the condition is true => empty + default value = default value

or if you use this in lots of place and to make it looks less ugly, try this

const textOrDefault = (defaultValue) => `concat(
    text(),
    substring(
        "${defaultValue}", 
        1, 
        number(not(text())) * string-length("${defaultValue}")
    )
)`
const template = {
    items: ['//items/item', textOrDefault('my default val')]
}

[tuananh]

the if a path is empty, return blank is just that if you use a empty string in template, the output will be blank. that's it. Mostly useful in case you want to fill them in with empty string instead of leaving them undefined.

[aaronjfox]

That makes sense. I'll be able to put that workaround to use. Thanks for the help.

[eeveefox]

Hello. how i can do someone like that, but with attribute. `

`

if i use it with start: ['insert/dbCommand/start_ute_command/ute_command/@call', textOrDefault('false')], i alwayse have 'false'

[tuananh]

@eeveefox it seems right. can you post the whole xml and your code snippet so that i can try

[eeveefox]

<root>
    <!--  Тестовые команды для семантических данных  -->
    <data table="TEST_LAYER_SEM" schema="TEST" id="TEST_LAYER_SEM" comment="TEST_LAYER_SEM">
        <select>
            <dbQuery idField="ID">
                <var name="FILTER" default="1=1"/>
                <start_ute_command>
                    <ute_command call="TEST.xml#TEST_PREV_COMMAND" />
                </start_ute_command>
                <query>
                    SELECT d.doc_version_id, d.doc_id id, d.next_doc_version_id, d.mime_type_id, d.storage_id, d.name,
                    d.descr, d.version_number, d.body, d.file_name, d.ctime, d.cuser_id, d.mtime, d.muser_id
                    FROM  web50.lib_doc_version_node d WHERE {FILTER} LIMIT 2
                </query>
            </dbQuery>
        </select>
    </data>
</root>

and

const recipeTemplate = {
            root: ["/root/data", {
                id: "@id",
                select: {
                    query: "select/dbQuery/query",
                    vars: ["select/dbQuery/var", {
                        name: "@name",
                        type: "@type",
                        default: "@default",
                        direction: "@direction"
                    }],
                    params: ["select/dbQuery/param",{
                        name: "@name",
                        type: "@type",
                        default: "@default"
                    }],
                    ute_commands: {
                        start: ['select/dbQuery/start_ute_command/ute_command/@call', textOrDefault('false')],
                        end: "select/dbQuery/end_ute_command/ute_command/@call"
                    }
                }
             }
            ]
        };

[tuananh]

@eeveefox ah because the textOrdefautl value only give you output of text() xpath function

notice what i change in the textOrDefault function and the template below.

you can try this

const { transform } = require('.')
const fs = require('fs')

;(async function main() {
    const textOrDefault = (defaultValue) => `concat(
        @call,
        substring(
            "${defaultValue}", 
            1, 
            number(not(@call)) * string-length("${defaultValue}")
        )
    )`

    const recipeTemplate = {
        root: ["/root/data", {
            id: "@id",
            select: {
                query: "select/dbQuery/query",
                vars: ["select/dbQuery/var", {
                    name: "@name",
                    type: "@type",
                    default: "@default",
                    direction: "@direction"
                }],
                params: ["select/dbQuery/param",{
                    name: "@name",
                    type: "@type",
                    default: "@default"
                }],
                ute_commands: {
                    start: ['select/dbQuery/start_ute_command/ute_command', textOrDefault('false')],
                    end: "select/dbQuery/end_ute_command/ute_command/@call"
                }
            }
         }
        ]
    };

    const xml = fs.readFileSync('input.xml', 'utf-8')
    const output = await transform(xml, recipeTemplate)
    console.log(JSON.stringify(output, null, 4));

})()

output

{
    "root": [
        {
            "id": "TEST_LAYER_SEM",
            "select": {
                "query": "\n                    SELECT d.doc_version_id, d.doc_id id, d.next_doc_version_id, d.mime_type_id, d.storage_id, d.name,\n                    d.descr, d.version_number, d.body, d.file_name, d.ctime, d.cuser_id, d.mtime, d.muser_id\n                    FROM  web50.lib_doc_version_node d WHERE {FILTER} LIMIT 2\n                ",
                "vars": [
                    {
                        "name": "FILTER",
                        "type": "",
                        "default": "1=1",
                        "direction": ""
                    }
                ],
                "params": [],
                "ute_commands": {
                    "start": [
                        "TEST.xml#TEST_PREV_COMMAND"
                    ],
                    "end": ""
                }
            }
        }
    ]
}

[eeveefox]

Thank you, I didn't figure it out myself. It's the right thing to do.