Search across everything

Find datasets, tables, and columns in one place

Search looks across datasets, schemas, tables, columns, and publishers from a single box. Use it when the thing you are looking for spans datasets — for example, “which tables have this column?”

Open search

Type a keyword into the search box in the header and press Enter. ⌘K (Ctrl+K on Windows) also moves the cursor into the box. It works the same way from any page.

Switch what you are searching for

The left sidebar lists what you can search for, each with a match count.

Tables: the tables themselves. This is the default
Columns: the columns inside tables
Datasets: a dataset’s own name, description, and keywords
Schemas: the groupings inside a dataset
Users / Organizations: the publishers of the datasets

The counts are for the keyword you typed, so you can tell at a glance when there is nothing under tables but several matches under columns. Switching targets keeps your keyword.

How keywords match

Each target matches against different fields.

Tables: the table’s name, title, description, and tags, plus the name of the dataset it belongs to
Columns: the column’s name, title, and description. Table titles are not matched. This target exists to answer “which tables have a population column”, so including columns that only matched through their table would defeat the purpose
Datasets: the dataset’s own name, title, description, and keywords

Column results list matches on the column name and title first; columns that matched only inside a description come after.

Separate keywords with spaces to require all of them.

Use “Dataset” in the left sidebar to keep only results from a particular dataset. The counts reflect the distribution after your keyword is applied, so you can see which dataset the results are concentrated in.

Find tables by their columns

When you are looking for tables, you can narrow them by the columns they hold. If you want population by town block, look for tables with a town block column and a population column. The keyword can stay empty.

Column conditions: type a condition and press Enter to add it. Only tables with a matching column for every condition remain, so entering both 町丁目 and 人口 leaves tables that have both. Conditions match column names, titles and descriptions
Key columns: narrow by what a table’s rows are identified by: prefecture, municipality, small area, grid square, corporate number or postal code. Selecting several keeps tables with any of them
Has location data: keep only tables with a geometry column or a latitude/longitude pair

Key columns are declared by the publisher (Declare key columns). A column condition that names a key, such as 町丁目, also matches columns declared with that key even when they are written as 町丁・字等 or KEY_CODE.

When nothing matches

If there are no results, any other target that does have matches is shown with its count and a link. Changing the level often finds what you were looking for.

Next steps

Explore datasets: browse from the dataset list
View table details: columns, statistics, and lineage
Connect: reach the data you found from external tools