For my Jekyll blog, I'm currently trying to add a feature that, for each page, processes data from an associated JSON file. I'm using the following structure (_data directory with a subdirectory comments-gh, which contains the JSON files):

./_data/
└── comments-gh
    ├── 14.json
    ├── 15.json
    ├── ...
    ├── ...
    ├── 76.json
    └── 77.json

The number that makes up the base name of each file corresponds to a numerical identifier that is defined in the YAML frontmatter of each page by variable pagement_id:

---
comment_id: 76
---

In theory this should allow me to access, for each page, the corresponding JSON, and then process it using something like this (or so I hope!):

{% assign comments = site.dataments-gh.pagement_id %}
{% for comment in comments %}

    (processing code)

{% endfor %}

As it turns out, this doesn't actually work at all! But it does work if I hard-code the comment_id value like this:

{% assign comments = site.dataments-gh.76 %}

So I suspect I'm doing something wrong with the way I'm using the pagement_id variable. I couldn't find any clear examples in the Jekyll docs (which mostly cover the case of 1 single data file), but after some searching I did find this thread, which covers a similar issue:

So, I adapted my code along the lines of "Option #1" (the author of that post explicitly mentions omitting the dot and putting the variable between square brackets):

{% assign comments = site.dataments-gh[pagement_id].items %}

But that didn't solve the issue either! I'm probably overlooking something very basic (I don't work with Liquid often).

Does anyone have any suggestions?

(BTW I double-checked that pagement_id is actually defined and has the expected value by writing its value to HTML output.)

For my Jekyll blog, I'm currently trying to add a feature that, for each page, processes data from an associated JSON file. I'm using the following structure (_data directory with a subdirectory comments-gh, which contains the JSON files):

./_data/
└── comments-gh
    ├── 14.json
    ├── 15.json
    ├── ...
    ├── ...
    ├── 76.json
    └── 77.json

The number that makes up the base name of each file corresponds to a numerical identifier that is defined in the YAML frontmatter of each page by variable pagement_id:

---
comment_id: 76
---

In theory this should allow me to access, for each page, the corresponding JSON, and then process it using something like this (or so I hope!):

{% assign comments = site.dataments-gh.pagement_id %}
{% for comment in comments %}

    (processing code)

{% endfor %}

As it turns out, this doesn't actually work at all! But it does work if I hard-code the comment_id value like this:

{% assign comments = site.dataments-gh.76 %}

So I suspect I'm doing something wrong with the way I'm using the pagement_id variable. I couldn't find any clear examples in the Jekyll docs (which mostly cover the case of 1 single data file), but after some searching I did find this thread, which covers a similar issue:

https://talk.jekyllrb/t/page-variable-name-in-for-loop-using-data-files/7522/7

So, I adapted my code along the lines of "Option #1" (the author of that post explicitly mentions omitting the dot and putting the variable between square brackets):

{% assign comments = site.dataments-gh[pagement_id].items %}

But that didn't solve the issue either! I'm probably overlooking something very basic (I don't work with Liquid often).

Does anyone have any suggestions?

(BTW I double-checked that pagement_id is actually defined and has the expected value by writing its value to HTML output.)

• jekyll
• liquid

Share Improve this question Follow asked Mar 3 at 19:18 johanjohan 87411 gold badge88 silver badges2121 bronze badges 0

Add a comment  | 

2 Answers 2

Sorted by: Reset to default 1

I think the .items part is wrong (it's not mentioned here). In your case, I think the syntax should be just:

{% assign comments = site.dataments-gh[pagement_id] %}

The combination of @andy-jackson's answer, and the comment by @mb21 about data types put me on the right track. It turns out that my problem was caused by the fact that the value of pagement_id is an integer, whereas it needs to be a string to work in the file reference.

Surprisingly (at least to me), Liquid doesn't appear to have any built-in tags to convert an integer to a string! After some searching I did find this workaround that appends two quotation marks to an integer variable, which then magically returns a string. This works, but it's a bit hacky for my taste.

Digging into the docs on Liquid variables I did find the "capture" tag, which captures the string inside two tags, and then assigns the result to another (string) variable. I applied this to pagement_id, and stored the result in a new variable comment_id. I then used that variable in the way suggested by @andy-jackson.

Here's what I ended up with:

{% capture comment_id %}{{ pagement_id }}{% endcapture %}
{% assign comments = site.dataments-gh[comment_id] %}

{% for comment in comments %}

    (processing code)

{% endfor %}

With these changes, the data file references work as expected on each page.