Question: How can data formatting be configured (such as prefix, number of decimal places to retain)?

Answer: The formatting of the current query values is determined by the "dataset" field's "data format" setting. Users can modify this in "Data preparation" - "Dataset" - "Edit field" - "Data format".
The default is "Auto", which intelligently adds thousand separators and retains 2 or 4 decimal places.

To change the default formatting, for example, to display 1.2324 as $1.22, simply modify the "Data format" configuration for the corresponding field.

Question: Why was chart A expected, but chart B was rendered instead?

Answer: If the fields (type and quantity) configured for the agent's data are not suitable and cannot support rendering the expected chart type, another chart type will be used for rendering. For detailed chart descriptions, see Common chart analysis.

• When the number of dimension and measure fields recalled by SQL does not meet expectations.
  For example: a pie chart cannot simultaneously display "Product", "Sales", and "Profit". This is because a pie chart can only accept one measure for size mapping, such as "Sales".

  Chart type	Dimension	Measure	Description
  Line chart	[1,10]	[1,10]	A line chart requires at least one dimension (as the X axis) and one measure (as the Y axis).
  Bar chart	[1,10]	[1,10]	A bar chart requires at least one dimension (as the X axis) and one measure (as the Y axis).
  Pie chart	[1,1]	[1,1]	A pie chart only accepts one dimension (as the name) and one measure (as the size).
  Metric card	[0,1]	[1,8]	A metric card can accept at most one dimension (as the core information) and multiple measures (as auxiliary information).
  Table	/	/	A table has no restrictions on dimensions or measures.
  Scatter plot	[0,100]	[2,4]	A scatter plot has no restrictions on dimensions. It requires at least two measures (as the X and Y axes), and up to four measures (the third as color, the fourth as size).
  China map	[1,1]	[1,1]	The China map only accepts one dimension (as province) and one metric (as color).
  Dual-axis chart	[1,1]	[2,10]	The dual-axis chart only accepts one dimension (as X axis) and at least two metrics (as left and right dual Y axes).
  Combination chart	[1,1]	[2,10]	The dual-axis chart only accepts one dimension (as X axis) and at least two metrics (as upper and lower dual Y axes).

• Due to differences in chart presentation, if the user does not explicitly specify the chart type, other chart types may be used for better results when there are too many or too few items to render.
  For example: if a pie chart displays 300 segments, the large amount of data makes the pie chart extremely difficult to read, so a pie chart will not be used in non-mandatory scenarios.

  Chart types	Recommended maximum rendering quantity	Note
  Line chart	366	The line chart is suitable for rendering 366 points (covering the scenario of one year of detailed data). When rendering more points, crowding becomes severe and readability is low, making tables preferable.
  Bar chart	181	The bar chart is suitable for rendering 181 bars (covering the scenario of half a year of detailed data). When rendering more bars, crowding becomes severe and readability is low, making tables preferable.
  Pie chart	181	The pie chart is suitable for rendering 181 sectors (covering the scenario of half a year of detailed data). When rendering more sectors, crowding becomes severe and readability is low, making tables preferable.
  Metric card	20	The metric card is suitable for rendering 20 metric blocks. When there are more metric blocks, key information is lost, readability is low, and tables are preferable.
  Scatter plot	1000	The scatter plot is suitable for rendering 1000 scatter points. When rendering more points, crowding becomes severe and readability is low, making tables preferable.
  China map	34	The China map displays up to 34 blocks (corresponding to the 34 primary administrative regions on the China map); there are no other applicable scenarios.
  Dual-axis chart	181	The dual-axis chart is suitable for rendering 181 points (covering the scenario of half a year of detailed data). When rendering more points, crowding becomes severe and readability is low, making tables preferable.
  Combination chart	181	The combination chart is suitable for rendering 181 bars (covering the scenario of half a year of detailed data). When rendering more bars, crowding becomes severe and readability is low, making tables preferable.
  Table	No upper limit	There is no limit to the number of items rendered in tables.