Returns the largest value of all array elements which satisfy one or several conditions. MAXIFS ({ 10, 20, 30 }, { 1, 2, 3 }, ">1", { "Eve", "Eve", "Bill" }, "Eve") MAXIFS ({ 10; 20; 30 }; { 1; 2; 3 }; ">1"; { "Eve"; "Eve"; "Bill" }; "Eve") returns 20, because that is the only element of the first array where both corresponding elements in the other arrays satisfy their conditions.

In the second array { 1, 2, 3 }{ 1; 2; 3 }, 2 and 3 satisfy the ">1"">1" condition, as they are both greater than 1. In the third array, { "Eve", "Eve", "Bill" }{ "Eve"; "Eve"; "Bill" }, the two first elements satisfy the "Eve""Eve" condition, requiring matching elements to be equal to "Eve". In other words, the second and third elements of the second array match their corresponding condition, while the first and second elements of the third array match their corresponding condition. As such, only the second element of the first array is considered, resulting in a return value of 20.

Use & to reference values that reside elsewhere. This formula only considers those elements of the first array which are greater than NumberValue :

MAXIFS works with multiple conditions, but it only requires a single condition. There is no function named MAXIF.

Writing conditions

MAXIFS supports the following operators: < (less than), <= (less than or equal to), > (greater than), >= (greater than or equal to), = (equality, case-insensitive for text), == (equality, case-sensitive for text), <> (inequality, case-insensitive for text) and != (inequality, case-sensitive for text).

If a condition parameter is set to "" , only the blank values of the corresponding array are considered. If it is set to "<>" or "!=" , all non-blank values are considered.

If the operator is left out, "=" is assumed. As a result, these formulas are equivalent:

The configured locale is taken into account when parsing numbers that are part of a condition text string, meaning that if the locale is German, a decimal comma (",") should be used as a decimal separator. Similarly, if the locale is Australian English, a decimal point (".") should be used. Thousands separators should not be used and negative numbers should use a leading minus sign ("-").

Specifying the condition using a formula fragment

The condition parameters can also be formula fragments. These formulas are equivalent:

Using a formula fragment gives you access to the full power of the formula locale. Consider this formula:

The formula above returns 30, because only array elements are considered which are greater than 1 and are odd numbers, which disqualifies 1 and 2, leaving only 3 (corresponding to 30 in the first array).

&& means logical "and," || means logical "or" and ! means logical negation.

This formula returns 30, as 1, 2 and 3 are all considered:

1, 2 and 3 are all considered due to the || operator, which causes all array elements to be considered which are either greater than 1 or are odd (1, 2 and 3 correspond to 10, 20 and 30 in the first array).

This formula returns 20, as only 2 is both greater than 1 and is "not odd" (even):

! in ! ISODD (Element) ! ISODD (Element) negates the value returned from ISODD, causing TRUE to become FALSE and vice versa.

The formula fragment is run once for every array element to be tested, and is expected to return TRUE if the element should be considered, and FALSE otherwise. To do its work, it has access to the values Element (the element of the corresponding array currently being tested), Index (the numeric position of the array element currently being tested, starting at 1) and Source , the corresponding array.

If MAXIFS is used with multiple conditions, either all or none may use formula fragments.

Using wildcards in the condition to match text flexibly

When a condition array contains text strings, the condition can use special characters to match multiple text strings. This formula returns 100, because both "car" and "cat" match the wildcard "ca?":

? matches any character. "ca?" matches not only "car" and "cat", but also "cab", "can" and "cap". However, it does not match "card", because ? matches exactly one character, nor does it match "ca".

* matches no character, a single character, or multiple characters. As such, "ca*" matches everything that "ca?" matches, but also "card", "ca" and "Canada" (matching is case-insensitive).

? and * can be combined in the same pattern. Colo*r ad*s are great? matches both "Color ads are great!" and "Colour adverts are great?".

Sometimes, the special characters ? and * need to be included as-is, and not have special meaning. To achieve that, precede each special character with a tilde ("~"). This formula returns 10, because only "ca*" in the second array matches "ca~*":

In other words, to search for a verbatim "*" character, write "~*". To search for a verbatim "?" character, write "~?". Finally, to search for a verbatim "~" character, write "~~".

Regular expressions are much more powerful than wildcards, but are also harder to learn. Use REGEXMATCH in a formula fragment to use regular expressions with MAXIFS.

This formula returns 10, because only "123" matches the "\d\d\d" regular expression (which only accepts exactly three numbers):

Refer to the documentation of REGEXMATCH for more information.

Using FILTER and MAX instead of MAXIFS

The work MAXIFS performs can be divided into two parts: filtering out the unwanted elements from the condition arrays and then returning the largest of the elements of the first array which correspond to the ones in the condition arrays which made the cut. As such, these formulas are equivalent:

The version using FILTER provides more flexibility, as the final step (determining the largest number through MAX) can be replaced by another function (like PRODUCT, which multiplies the remaining numbers together), independent of the filtering logic.

Of course, MAXIFS is normally used with multiple conditions. FILTER also supports multiple conditions using the && or * operator. These formulas are equivalent:

For even more flexibility, consider replacing MAX with REDUCE. Refer to the example below for more information.