New Analysis of $B\to K\pi$ data with Pauli blocking CP violation in $B^o$ decays, not in $B^{\pm }$ SU(3) use of $B\to \pi \pi$ data invalid for $B\to K\pi$

New data analysis with Pauli blocking explains observation of CP violation in $B^o\to K\pi$ decays, absence in $B^{\pm }\to K\pi$ decays and gives new predictions agreeing with experiment. Branching ratio data show pure I=1/2 amplitude predicted by pure penguin transitions for separate relations within charged and neutral B decays, but strong violation of penguin I=1/2 isospin relation between charged and neutral decays. $B(B^o \rightarrow K^+ \pi^-) - 2B(B^o\rightarrow K^o \pi^o) = %(19.4 \pm 0.6)- 2\cdot (9.4 \pm 0.6) = (0.6 \pm 1.3)\cdot 10^-6 \approx 0$ $2B(B^+ \rightarrow K^+ \pi^o) - B(B^+ \rightarrow K^o \pi^+) = %(25.8\pm 1.2) - (23.1 \pm 1.0) = (2.7 \pm 1.6)\cdot 10^-6 \approx 0$ ${{\tau^o}\over{\tau^+}}\cdot 2B(B^+ \rightarrow K^+ \pi^o) - B(B^o \rightarrow K^+ \pi^-) = (4.7 \pm 0.82)\cdot 10^-6 \not= 0$ This contrast between pure I=1/2 observed in individual charged and neutral final states and I=1/2 violation in relations between them is unexpected in previous treatments. Pauli blocking predicts this contrast by noting that two identical $u$ quarks in a relative s-wave are Pauli blocked. Tree diagram $\bar b\rightarrow \bar s u \bar u$for$\bar b$producing$u$ quark at weak vertex and tree-penguin interference producing CP violation are Pauli suppressed for $B^{+}$ decays with identical $u$ quark spectator. No suppression in $B^o$ decays with spectator $d$ quark. $B\to K\pi$ data analysis discards all amplitudes containing two identical $u$ quarks in final state and predicts observed isospin relations by using only Pauli-favored amplitudes $B_u\to \bar{s}du\bar{d}$ and $B_d\to \bar{s}ud\bar{u}$. Pauli-favored transitions explain dependence on flavor of spectator quark which does not participate in the weak interaction.