![]() |
Metamath Proof Explorer |
< Previous
Next >
Nearby theorems |
|
Mirrors > Home > MPE Home > Th. List > difsnid | Structured version Visualization version GIF version |
Description: If we remove a single element from a class then put it back in, we end up with the original class. (Contributed by NM, 2-Oct-2006.) |
Ref | Expression |
---|---|
difsnid | ⊢ (𝐵 ∈ 𝐴 → ((𝐴 ∖ {𝐵}) ∪ {𝐵}) = 𝐴) |
Step | Hyp | Ref | Expression |
---|---|---|---|
1 | snssi 4833 | . 2 ⊢ (𝐵 ∈ 𝐴 → {𝐵} ⊆ 𝐴) | |
2 | undifr 4506 | . 2 ⊢ ({𝐵} ⊆ 𝐴 ↔ ((𝐴 ∖ {𝐵}) ∪ {𝐵}) = 𝐴) | |
3 | 1, 2 | sylib 218 | 1 ⊢ (𝐵 ∈ 𝐴 → ((𝐴 ∖ {𝐵}) ∪ {𝐵}) = 𝐴) |
Colors of variables: wff setvar class |
Syntax hints: → wi 4 = wceq 1537 ∈ wcel 2108 ∖ cdif 3973 ∪ cun 3974 ⊆ wss 3976 {csn 4648 |
This theorem was proved from axioms: ax-mp 5 ax-1 6 ax-2 7 ax-3 8 ax-gen 1793 ax-4 1807 ax-5 1909 ax-6 1967 ax-7 2007 ax-8 2110 ax-9 2118 ax-ext 2711 |
This theorem depends on definitions: df-bi 207 df-an 396 df-or 847 df-tru 1540 df-fal 1550 df-ex 1778 df-sb 2065 df-clab 2718 df-cleq 2732 df-clel 2819 df-rab 3444 df-v 3490 df-dif 3979 df-un 3981 df-in 3983 df-ss 3993 df-nul 4353 df-sn 4649 |
This theorem is referenced by: fnsnsplit 7218 fsnunf2 7220 difsnexi 7796 difsnen 9119 enfixsn 9147 pssnn 9234 phplem2OLD 9281 dif1ennnALT 9339 frfi 9349 xpfiOLD 9387 dif1card 10079 hashgt23el 14473 hashfun 14486 fprodfvdvdsd 16382 prmdvdsprmo 17089 mreexexlem4d 17705 symgextf1 19463 symgextfo 19464 symgfixf1 19479 gsumdifsnd 20003 gsummgp0 20341 islindf4 21881 scmatf1 22558 gsummatr01 22686 tdeglem4 26119 finsumvtxdg2sstep 29585 dfconngr1 30220 satfv1lem 35330 bj-raldifsn 37066 lindsadd 37573 lindsenlbs 37575 poimirlem25 37605 poimirlem27 37607 hdmap14lem4a 41828 hdmap14lem13 41837 supxrmnf2 45348 infxrpnf2 45378 fsumnncl 45493 hoidmv1lelem2 46513 gsumdifsndf 47904 mgpsumunsn 48086 |
Copyright terms: Public domain | W3C validator |