Tables

Most patent research starts with far more documents than anyone can read. A search for prior art or a competitor's portfolio easily returns thousands of patents, and the actual work is to narrow them down to the few that matter. Tables are built for exactly that: you start with everything and cut with filters and sorting until only the rows worth opening remain.

A table is a grid of records inside a research. Each row is one record and each table lists one kind of record, either patents or products. A dashboard shows the same data the other way around, as charts of the whole set, so in general you use a table to triage and act and a dashboard to see the overall picture.

Here is an example of a patents table. The first columns carry the patent's own fields, and every product under review adds a column of mapping scores colored by risk:

One look tells you that two patents, Toyota's and QuantumScape's, fully match the Aurora column, a few more score high and the rest sit low enough to wait. The coloring turns thousands of scores into something you can scan in seconds.

Rows and columns#

Every row is one record, a patent or a product, never a mix. Columns come in three kinds:

• The identifier column names the record and is always present.
• Attribute columns carry the record's attributes, system and custom alike.
• Mapping columns show how the record scores against each record on the other side.

Identifier column#

The left-most column of a table is reserved for the identifier. On a desktop it stays pinned in place while you scroll the other columns sideways, and it is the one column you cannot hide. For patents the identifier is the publication number, for products the name.

Attribute columns#

Attribute columns show a record's attributes: the system fields Patalyze fills in from the source and the custom fields you or an agent add. On a patents table that means titles, applicants, dates, classifications and status next to your own columns, e.g. a reviewer or a priority. A products table has the websites built in, and beside it whatever the research agent created and filled while it browsed, such as categories and manufacturers, plus anything you add yourself. Show the ones you need and hide the others. When the field you want does not exist yet, the plus button at the end of the columns is where you create it.

Mapping columns#

Mapping columns are the reason a table exists in patent research. Each one stands for a record on the other side: on a patents table, every product you are checking against adds a column. Where a patent and a product have been compared, the cell at their intersection holds the mapping between them.

A patent has many claims, but the cell leads with the strongest match: the claim that scores highest against the product, shown as a chip with its claim number and score. Other notable claims sit beside it and the remainder is collapsed into a count. A glance down the column tells you which patents come closest to reading on your product. Click a cell and choose "Open in panel" to read the mapping beside the table, or "Open as page" to walk through it claim by claim.

The color is the score's risk band. Every cell falls into one of four bands, from critical down to low, so the dangerous matches stand out at a glance. The mappings section of the data model explains each band and the percentage range behind it.

The cell reads the same from either side. A products table puts your products in the rows and patents in the columns, but a critical match is critical whichever way you came at it.

Filtering#

Filtering is how you narrow the field. Pick a column, choose an operator and the rows that fail the test drop away. Every option carries a count, so you can see how many rows a choice keeps before you commit. Add a filter from the toolbar, or open any column's header menu and choose "Filter by" to filter by that column directly.

The most useful move is filtering a mapping column to the critical and high bands. Thousands of rows collapse to the few patents whose claims actually threaten your product, and those are the ones worth opening.

Layer a second filter on top to sharpen the shortlist, e.g. a single applicant or in-force patents only. Once the list is right, tick the rows you want and a footer appears with four actions:

• Ask agent opens a research agent with exactly these records in hand, e.g. to fill an attribute for them or to chart them against a product. The same button appears when you select cells instead of rows.
• Copy puts the publication numbers or product names on the clipboard.
• Download exports the selection as CSV or Excel, columns included.
• Delete removes the selected records from the research, e.g. to clear out rows a filter proved irrelevant.

To download the whole table rather than a selection, use "Download as CSV" or "Download as Excel" from the page menu.

Sorting#

Where a filter removes rows, sorting reorders the ones that remain. Open a column's header menu and choose "Sort ascending" or "Sort descending". A small arrow on the header marks the active direction, and "Clear sorting" removes it again. Sort a mapping column from high to low and the closest matches float to the top before you have set a single filter.

Searching#

Search jumps straight to a record you already have in mind. Type a publication number, a title or an applicant into the search box and the table keeps only what matches the text. A filter narrows by a rule and search narrows by free text, so when all you want is to pull up one known patent, search is faster than building a filter.

Visualizing a column as a chart#

Any column can become a chart without leaving the table. Open the column's header menu and choose "Visualize as chart". A preview appears, binned to suit the column: a status column as a breakdown, a date as a trend, an applicant as a ranking. The preview respects your filters, so it charts exactly the rows you have narrowed down to, and those filters travel with it. "Add to dashboard" keeps the chart on a dashboard, or "Create dashboard" starts a new one if the research has none yet.

Arranging columns#

Not every column needs to be on screen at once. Hide one straight from its header menu with "Don't show on this table", or open the display settings to manage them all: switch the row height between "Compact" and "Expanded", drag columns into a new order and add any hidden column back when you need it again.

A few more adjustments live on the header itself. Drag the edge between two headers to set a column's width and double-click that edge to reset it. Drag a header sideways to move the column, or use "Move left" and "Move right" in its menu to nudge it one step at a time. "Select all cells" selects the whole column, e.g. to copy it or to ask an agent about it.

Custom attributes lead the menu with "Edit attribute", which changes the attribute itself rather than this one table. "Delete" sits at the bottom of that form, and deleting there removes the column from every page and takes its values with it.

Adding columns#

The plus button at the end of the columns opens one search box over everything you might add: the attributes that exist but are hidden, and the records on the other side that would become mapping columns. Type a few letters and both kinds filter together, so adding a product column and un-hiding a date column is the same gesture. Beneath the list sits "Create new attribute" to define a field that does not exist yet.

A record that is not in the research at all has to arrive first, through the Add data menu or a research agent. Once it is there, it shows up in this list like any other. While a product search is running, the table header shows how many sources have been scanned as new products arrive in the rows beneath.