Units of the p-adic integers #
Complements to Mathlib's PadicInt API on the units of ℤ_p: 1 + x is a unit whenever
p ∣ x, because ℤ_p is a local ring whose maximal ideal is pℤ_p. This is the criterion
that makes 1 + p^f ℤ_p a subgroup of ℤ_pˣ.
The unit group ℤ_pˣ is a profinite group: it is the closed unit sphere of the compact
totally disconnected space ℤ_p, and the topology of the units is the subspace topology because
ℤ_p is a complete normed ring. The instances CompactSpace ℤ_[p]ˣ and
TotallyDisconnectedSpace ℤ_[p]ˣ are recorded here.
Two consequences of the ultrametric divisibility in ℤ_p are recorded as well: an element
divides every element of no larger norm, so that a finite family of p-adic integers is a common
multiple q • w of a family w with a coordinate equal to 1, namely at an index of maximal norm.
This is the shape in which the exponent vector of a relator of a free pro-p group is read.
Main results #
PadicInt.isUnit_one_add_of_dvd:1 + xis a unit ofℤ_[p]wheneverp ∣ x.PadicInt.isUnit_two:2is a unit ofℤ_[p]for oddp.PadicInt.pow_p_dvd_natCast_iff:p ^ ndivides a natural number inℤ_[p]exactly when it does inℕ.PadicInt.dvd_of_norm_le: inℤ_[p],y ∣ xwhenever‖x‖ ≤ ‖y‖.PadicInt.exists_apply_eq_one_and_eq_smul: a finite family inℤ_[p]isq • wwithw i₀ = 1at some indexi₀.PadicInt.units_neg_one_ne_one:-1 ≠ 1inℤ_[p]ˣ.PadicInt.range_units_val: the units ofℤ_[p]are the elements of norm1.Padic.exists_eq_zpow_valuation_mul: every nonzerox : ℚ_[p]isp ^ v(x)times a unit ofℤ_[p].PadicInt.compactSpace_units,PadicInt.totallyDisconnectedSpace_units:ℤ_[p]ˣis a profinite group.
ℤ_pˣ is compact: it is the closed unit sphere of the compact space ℤ_p, and the topology
on the units of the complete normed ring ℤ_p is the subspace topology.
ℤ_pˣ is totally disconnected, as a subspace of the ultrametric space ℤ_p.
A finite family of p-adic integers is a multiple of a family with a coordinate 1: for
v : ι → ℤ_p with ι finite and nonempty there are an index i₀, a scalar q and a family w
with w i₀ = 1 and v = q • w. One may take i₀ of maximal norm and q = v i₀, which then
divides every coordinate.