12.1: Accessing Individual Elements

We can use the len() function, which we’ve already learned two uses for, in yet a third way: to ascertain the number of key/value pairs in a series. Using the Figure 11.1.2 example:

Code \(\PageIndex{1}\) (Python):

print(len(alter_egos))

| 4

Accessing the value for a given key uses exactly the same syntax that NumPy arrays used (boxies), except with the key in place of the numeric index:

Code \(\PageIndex{2}\) (Python):

superhero = alter_egos['Peter']

print("Pssst...Peter is really {}.".format(superhero))

| Pssst...Peter is really Spidey.

This is why it’s important that the keys of an associative array be unique. If we type “alter_egos['Peter'],” we need to get back one well-defined answer, not an ambiguous set of alternatives.¹ The values, on the other hand, may very well not be unique.

To overwrite the value for a key with a new value, just treat it as a variable and go:

Code \(\PageIndex{3}\) (Python):

alter_egos['Bruce'] = 'Batman'

print(alter_egos)

| Bruce Batman

| Peter Spidey

| Tony Iron Man

| Thor Thor

| dtype: object

This same syntax works for adding an entirely new key/value pair as well:

Code \(\PageIndex{4}\) (Python):

alter_egos['Diana'] = 'Wonder Woman'

print(alter_egos)

| Bruce Batman

| Peter Spidey

| Tony Iron Man

| Thor Thor

| Diana Wonder Woman

| dtype: object

It’s just like with ordinary variables, if you think about it. Saying “x=5” overwrites the current value of x if there already is an x, otherwise it creates a new variable x with that value.

Finally, to outright remove a key/value pair, you use the del operator:

Code \(\PageIndex{5}\) (Python):

del alter_egos['Tony'] ]

print(alter_egos)

| Bruce Batman

| Peter Spidey

| Thor Thor

| Diana Wonder Woman

| dtype: object

Bye bye, Iron Man.

Don’t get mad when I tell you that all of the above operations work in place on the Series, which is very different than some of the “return a modified copy” style we’ve seen recently. Hence all of these attempts are wrong:

Code \(\PageIndex{6}\) (Python):

alter_egos = del alter_egos['Tony'] <--- WRONG!

alter_egos = alter_egos['Bruce'] = 'Batman' <--- WRONG!

alter_egos = alter_egos['Diana'] = 'Wonder Woman' <--- WRONG!

You don’t “change a value and get a new Series”; you just “change it.”

Accessing by Position

One slightly weird thing you can do with a Pandas Series is ignore the key (index) altogether and instead use the number of the key/value pair to specify what value you want. This gives me the heebie-jeebies, because as I explained back on p. 57, there really isn’t any meaningful “order” to the key/value pairs of an associative array. In true All Things To All People™ fashion, however, Pandas lets you do this.

Accessing a Value by Position

You can ask for the value of (say) “the second” superhero. To do so, you use the bizarrely-named .iloc syntax:

Code \(\PageIndex{7}\) (Python):

a_hero = alter_egos.iloc[1]

print(a_hero)

| Spidey

This is occasionally useful, so I mention it for completeness. The .iloc numbers start with 0 (not 1) as is true throughout Python.

Accessing a Key by Position

Similarly, you can get the key (as opposed to the value) of the key/value pair at a particular position. To ask for the key of “the second” superhero, you use the .index syntax:

Code \(\PageIndex{8}\) (Python):

a_secret_hero = alter_egos.index[1]

print(a_secret_hero)

| Peter